XformAI-india/qwen-1.7b-coder
TEXT GENERATIONConcurrency Cost:1Model Size:2BQuant:BF16Ctx Length:32kPublished:May 2, 2025License:mitArchitecture:Transformer0.0K Open Weights Warm
XformAI-india/qwen-1.7b-coder is a 1.7 billion parameter Transformer decoder model, fine-tuned by XformAI from Qwen3 1.7B. This instruction-tuned model specializes in generating and completing code, particularly for Python, JavaScript, and Bash. It is optimized for low-latency inference and designed for use in developer tools, AI agents, and backend toolchains.
Loading preview...
XformAI-india/qwen-1.7b-coder: A Specialized Code Generation Model
qwen-1.7b-coder is a 1.7 billion parameter Transformer decoder model, fine-tuned by XformAI from the Qwen3 1.7B base model. Released in May 2025 under an MIT License, this model is specifically engineered for robust code generation across multiple programming languages.
Key Capabilities
- Code Generation: Purpose-built to generate and complete code snippets, focusing on Python, JavaScript, and Bash.
- Instruction-Tuned: Fine-tuned on a dedicated code corpus using LoRA (8-bit) + PEFT techniques for instruction-based code generation.
- Low-Latency Inference: Designed for efficient operation, making it suitable for real-time applications and edge devices.
- Training Precision: Utilizes bfloat16 precision during its 3-epoch fine-tuning process.
Good For
- Developer Tools: Ideal for creating VSCode-like autocomplete agents and on-device copilots.
- AI Agents: Enhances AI agents with reliable code generation capabilities.
- Backend Toolchains: Supports backend logic and API template generation.
- Chatbots: Can power code-aware chatbots and shell command assistants.