How to use from
llama.cpp
Install from brew
brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf animeshkundu/cmd-correct
# Run inference directly in the terminal:
llama-cli -hf animeshkundu/cmd-correct
Install from WinGet (Windows)
winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf animeshkundu/cmd-correct
# Run inference directly in the terminal:
llama-cli -hf animeshkundu/cmd-correct
Use pre-built binary
# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf animeshkundu/cmd-correct
# Run inference directly in the terminal:
./llama-cli -hf animeshkundu/cmd-correct
Build from source code
git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf animeshkundu/cmd-correct
# Run inference directly in the terminal:
./build/bin/llama-cli -hf animeshkundu/cmd-correct
Use Docker
docker model run hf.co/animeshkundu/cmd-correct
Quick Links

cmd-correct

Shell command correction models for the fix CLI. Fine-tuned Qwen3 models that correct typos, fix case errors, and translate natural language to shell commands.

Available Models

Model Size Base Description
qwen3-correct-0.6B.gguf 378 MB Qwen3-0.6B Fast, lightweight (default)
qwen3-correct-1.7B.gguf ~1 GB Qwen3-1.7B Higher quality, more capable

Quick Start

Install the CLI

macOS (Homebrew):

brew tap animeshkundu/fix
brew install fix

From Source (Rust):

cargo install --git https://github.com/animeshkundu/fix --features metal  # macOS
cargo install --git https://github.com/animeshkundu/fix --features cuda   # Linux/Windows with NVIDIA
cargo install --git https://github.com/animeshkundu/fix                   # CPU only

Usage

# Basic correction
fix gti status
# Output: git status

# With shell override
fix --shell powershell Get-ChidItem
# Output: Get-ChildItem

# Switch to larger model
fix --use-model qwen3-correct-1.7B

# List available models
fix --list-models

Supported Shells

bash, zsh, fish, powershell, cmd, tcsh

Model Details

Training

  • Method: LoRA fine-tuning with MLX on Apple Silicon
  • Dataset: 135K synthetic examples (typos, case errors, natural language)
  • Format: ChatML with <|im_start|> / <|im_end|> tokens
  • Quantization: Q4_K_M with importance matrix

Prompt Template

<|im_start|>system
You are a shell command corrector for {shell}. Output only the corrected command.<|im_end|>
<|im_start|>user
{incorrect_command}<|im_end|>
<|im_start|>assistant

Example Corrections

Input Output Type
gti status git status typo
git remote -V git remote -v case
dcokre ps docker ps typo
I need to list docker images docker images natural language

Links

License

Apache 2.0

Downloads last month
4
GGUF
Model size
0.6B params
Architecture
qwen3
Hardware compatibility
Log In to add your hardware

We're not able to determine the quantization variants.

Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for animeshkundu/cmd-correct

Finetuned
Qwen/Qwen3-0.6B
Quantized
(301)
this model