XXXiong/ChatHLS-HLSFixer

TEXT GENERATIONConcurrency Cost:1Model Size:14.8BQuant:FP8Ctx Length:32kPublished:Apr 20, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Cold

XXXiong/ChatHLS-HLSFixer is a 14.8 billion parameter language model, fine-tuned from Qwen/Qwen2.5-14B-Instruct, specifically designed for High-Level Synthesis (HLS) C/C++ error correction. It functions as the core analysis agent within the ChatHLS framework, specializing in debugging HLS-incompatible code. This model excels at identifying, analyzing, and proposing fixes for HLS-related bugs based on compiler error logs and code analysis. Its primary use case is automating the debugging process for HLS C/C++ programs.

Loading preview...

ChatHLS-HLSFixer: Specialized HLS C/C++ Debugging Model

ChatHLS-HLSFixer is a 14.8 billion parameter language model, fine-tuned from Qwen/Qwen2.5-14B-Instruct, developed by XXXiong. It is specifically engineered to address High-Level Synthesis (HLS) C/C++ error correction, serving as a critical component within the broader ChatHLS framework.

Key Capabilities

  • HLS Error Identification: Pinpoints erroneous code lines in HLS C/C++ programs.
  • Root Cause Analysis: Explains why specific HLS-incompatible bugs occur, leveraging compiler error logs and code differences.
  • Automated Fix Proposals: Generates precise code modifications to resolve identified HLS errors.
  • Chain of Thought (CoT) Debugging: Employs a structured, sequential reasoning process to debug, including hypothesis formation and verification.
  • Structured Output: Provides bug analysis in a standardized JSON format, detailing erroneous lines, reasons, and modification actions.

Good For

  • Automating HLS Debugging: Ideal for developers and researchers working with HLS who need to quickly identify and fix C/C++ code issues that prevent successful synthesis.
  • Integrating into HLS Workflows: Designed to be the core analysis agent in systems like ChatHLS, streamlining the design automation and optimization process.
  • Educational Purposes: Can assist in understanding common HLS C/C++ pitfalls and their solutions through its detailed analysis and CoT reasoning.

This model is particularly useful for tasks requiring deep understanding of HLS constraints and compiler feedback to generate actionable debugging instructions.

Finetunes
1 models