CLI_model / README.md
Maarij-Aqeel's picture
Update README.md
5da83f8 verified
metadata
tags:
  - code
  - code-generation
  - cli
  - bash
  - python
  - terminal
  - automation
  - transformers
  - lora
  - fine-tuned
language:
  - en
license: mit
datasets:
  - custom
base_model: deepseek-ai/deepseek-coder-v2-lite-base
model_type: causal-lm
library_name: transformers
pipeline_tag: text-generation

🛠️ cli-model-deepseek-v2-lite-finetuned

A fine-tuned version of DeepSeek-Coder-V2-Lite-Base specialized in generating and understanding command-line interface (CLI) tools, automation scripts, and terminal-based workflows.


🧠 Model Details

Field Value
Base Model DeepSeek-Coder-V2-Lite-Base
Fine-tuned on Custom CLI-focused code dataset
Objective Next-token prediction / code generation
Domain CLI tools, Bash scripts, Python CLI apps
Framework HuggingFace Transformers (PyTorch)
Finetuning Method LoRA (Low-Rank Adaptation)
Hardware 2× Tesla T4 GPUs
Training Duration ~5 hours

📂 Training Data

The model was trained on a custom dataset containing real-world command-line and scripting examples:

  • Shell scripts (.sh)
  • Python CLI utilities
  • Terminal workflows (e.g., Git, curl)
  • API wrappers and automation scripts
  • Dockerfiles and Makefiles

Format: JSONL with {"prompt": ..., "completion": ...} pairs


🎯 Use Cases

This model is designed for developers who:

  • Want to generate CLI-based tools and scaffolds
  • Need help with automating tasks using Bash or Python
  • Build terminal wrappers around REST APIs
  • Create documentation and examples for terminal apps