|
|
--- |
|
|
license: apache-2.0 |
|
|
base_model: Qwen/Qwen3-1.7B |
|
|
tags: |
|
|
- lora |
|
|
- zsh |
|
|
- cli |
|
|
- autocomplete |
|
|
- command-line |
|
|
- peft |
|
|
--- |
|
|
|
|
|
# duoyuncloud/zsh-assistant-lora |
|
|
|
|
|
This is a LoRA (Low-Rank Adaptation) adapter for [Qwen/Qwen3-1.7B](https://huggingface.co/Qwen/Qwen3-1.7B) fine-tuned for Zsh CLI command autocomplete. |
|
|
|
|
|
## Model Details |
|
|
|
|
|
### Base Model |
|
|
- **Base Model:** [Qwen/Qwen3-1.7B](https://huggingface.co/Qwen/Qwen3-1.7B) |
|
|
- **Model Type:** Causal Language Model |
|
|
- **Task:** CLI Command Completion |
|
|
|
|
|
### LoRA Configuration |
|
|
- **LoRA Rank (r):** 8 |
|
|
- **LoRA Alpha:** 16 |
|
|
- **LoRA Dropout:** 0.1 |
|
|
- **Target Modules:** k_proj, q_proj, o_proj, gate_proj, v_proj, down_proj, up_proj |
|
|
|
|
|
### Training Configuration |
|
|
- **Epochs:** 3 |
|
|
- **Learning Rate:** 0.0002 |
|
|
- **Batch Size:** 1 |
|
|
- **Training Data:** 277 CLI command completion samples (Git, Docker, NPM, Python, Kubernetes, etc.) |
|
|
|
|
|
## Usage |
|
|
|
|
|
### Using with PEFT |
|
|
|
|
|
```python |
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer |
|
|
from peft import PeftModel |
|
|
|
|
|
# Load base model |
|
|
base_model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3-1.7B") |
|
|
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen3-1.7B") |
|
|
|
|
|
# Load LoRA adapter |
|
|
model = PeftModel.from_pretrained(base_model, "duoyuncloud/zsh-assistant-lora") |
|
|
|
|
|
# Use for inference |
|
|
prompt = "Input: git comm\nOutput:" |
|
|
inputs = tokenizer(prompt, return_tensors="pt") |
|
|
outputs = model.generate(**inputs, max_new_tokens=50) |
|
|
result = tokenizer.decode(outputs[0], skip_special_tokens=True) |
|
|
print(result) |
|
|
``` |
|
|
|
|
|
### Using with Ollama |
|
|
|
|
|
This adapter is designed to work with the [zsh-llm-cli-autocomplete-tool](https://github.com/duoyuncloud/zsh-llm-cli-autocomplete-tool) project. |
|
|
|
|
|
After importing to Ollama, use it as: |
|
|
```bash |
|
|
ollama run zsh-assistant |
|
|
``` |
|
|
|
|
|
## Training Data |
|
|
|
|
|
The model was fine-tuned on 277 command completion pairs covering: |
|
|
- Git commands (status, add, commit, push, pull, etc.) |
|
|
- Docker commands (run, build, ps, exec, etc.) |
|
|
- NPM/Node commands (install, run, start, etc.) |
|
|
- Python commands (-m, -c, pip, etc.) |
|
|
- Kubernetes commands (get, apply, delete, etc.) |
|
|
- System commands (ls, cd, mkdir, etc.) |
|
|
|
|
|
## Limitations |
|
|
|
|
|
- This adapter is specifically fine-tuned for CLI command completion tasks |
|
|
- Performance may vary for other use cases |
|
|
- The base model's limitations also apply |
|
|
|
|
|
## Citation |
|
|
|
|
|
If you use this adapter, please cite: |
|
|
|
|
|
```bibtex |
|
|
@misc{duoyuncloud_zsh_assistant_lora}, |
|
|
title={Zsh CLI Autocomplete LoRA Adapter}, |
|
|
author={Your Name}, |
|
|
year={2024}, |
|
|
publisher={Hugging Face}, |
|
|
howpublished={\url{https://huggingface.co/duoyuncloud/zsh-assistant-lora}} |
|
|
} |
|
|
``` |
|
|
|
|
|
## License |
|
|
|
|
|
This adapter inherits the license from the base model [Qwen/Qwen3-1.7B](https://huggingface.co/Qwen/Qwen3-1.7B). |
|
|
|
