|
|
--- |
|
|
license: apache-2.0 |
|
|
datasets: |
|
|
- Phonepadith/laos-long-content |
|
|
language: |
|
|
- lo |
|
|
base_model: |
|
|
- Qwen/Qwen2.5-VL-7B-Instruct |
|
|
new_version: Phonepadith/Laollm |
|
|
pipeline_tag: text-generation |
|
|
library_name: fastai |
|
|
--- |
|
|
|
|
|
# 🧠 AIDC LLM ສຳລັບພາສາລາວ - ແລະ ສະຫລຸບຂໍ້ມູນ — Fine-tuned on AIDC-5K (Q8 GGUF) |
|
|
|
|
|
This is a **Lao language chat model** based on [`Qwen/Qwen1.5-7B`](https://huggingface.co/Qwen/Qwen1.5-7B), fine-tuned on the **AIDC 5K Lao prompt-completion dataset**, and exported to **GGUF format** with **Q8_0 quantization** for fast inference using `llama.cpp`, `LM Studio`, or `Ollama`. |
|
|
|
|
|
--- |
|
|
|
|
|
## 🧾 Model Summary |
|
|
|
|
|
| Feature | Details | |
|
|
|----------------------|----------------------------------------------| |
|
|
| Base Model | [Qwen1.5-7B](https://huggingface.co/Qwen/Qwen1.5-7B) | |
|
|
| Fine-tuned By | [Phonepadith Phoummavong](https://huggingface.co/Phonepadith) | |
|
|
| Language | Lao (lo) | |
|
|
| Dataset | AIDC 5K Lao Prompt-Completion Dataset | |
|
|
| Quantization | Q8_0 (8-bit, GGUF) | |
|
|
| Format | GGUF | |
|
|
| File Name | `seallm-7b-lao-finetuned-lora-q8.gguf` | |
|
|
| Size (est.) | ~3–5 GB | |
|
|
| License | apache-2.0 | |
|
|
|
|
|
--- |
|
|
|
|
|
## 💡 Use Cases |
|
|
|
|
|
- Lao chatbots or digital assistants |
|
|
- Cultural and educational tools in Lao |
|
|
- Research on low-resource language modeling |
|
|
- Lao-native prompt generation and dialogue completion |
|
|
|
|
|
--- |
|
|
|
|
|
## 📥 How to Use |
|
|
|
|
|
|
|
|
### 🔸 LM Studio |
|
|
|
|
|
1. Download the `.gguf` model file. |
|
|
2. Open **LM Studio**. |
|
|
3. Click **"Add Local Model"**. |
|
|
4. Load `aidc-laollm-5k-based-qwen-Q8.gguf`. |
|
|
|
|
|
### 🔸 Ollama |
|
|
|
|
|
1. Place the `.gguf` file in your models directory. |
|
|
2. Create a `Modelfile`: |
|
|
```Dockerfile |
|
|
FROM qwen:7b |
|
|
PARAMETER quantization=q8_0 |
|
|
|
|
|
|
