micymike/codemate-qwen-1.5B-8k

Hugging Face
TEXT GENERATIONConcurrent Unit Cost:1Model Size:1.5BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Jun 20, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Warm

micymike/codemate-qwen-1.5B-8k is a 1.5 billion parameter Qwen-based conversational coding assistant, developed by micymike. It features an extended 8,192-token context window and resolves prompt repetition issues through targeted fine-tuning. This model specializes in multi-turn code repair, bug isolation, and providing production-ready solutions, particularly for Python and ReactJS.

Loading preview...

CodeMate-Qwen-1.5B-8k Overview

CodeMate-Qwen-1.5B-8k is a 1.5 billion parameter, multi-turn conversational coding assistant, building upon the micymike/codemate-qwen-1.5B base. This iteration significantly enhances its utility by addressing key limitations of its predecessor.

Key Capabilities and Improvements

  • Extended Context Window: The model's context window has been substantially upgraded from 2,048 tokens to 8,192 tokens. This allows it to process larger codebases and maintain conversational memory across extensive debugging sessions.
  • Prompt Echoing Fixed: Through Supervised Fine-Tuning (SFT) with completion_only_loss=True using Hugging Face TRL, the model no longer repeats user instructions, ensuring cleaner and more direct responses.
  • Bug-Fixing Specialization: It has been fine-tuned on multi-turn code repair paths (specifically iamtarun/code_instructions_120k_alpaca), enabling it to effectively isolate and explain programming syntax bugs, and offer practical solutions. It retains strong baseline knowledge in Python and ReactJS.

Usage and Format

The model utilizes the standard ChatML prompt template. Users should structure their inference payloads with <|im_start|>user and <|im_start|>assistant tags to ensure optimal performance and prevent parsing issues.

Good For

  • Multi-turn code debugging: Its extended context and specialized training make it ideal for interactive bug identification and resolution.
  • Code repair and explanation: Providing detailed explanations for syntax errors and suggesting production-ready fixes.
  • Assisting with Python and ReactJS development: Leveraging its foundational knowledge in these languages for coding tasks.

Finetunes
2 models