Downloads · 30 days
0
megaketu555/kashmiri-nastaliq-llm
kashmiri-nastaliq-llm is a text generation model from megaketu555. Use it when you need the model to write or continue text. The card lists the license as apache-2.0.
A conversational LLM fine-tuned for Kashmiri language in Nastaliq script (نَستَعلیٖق).
Downloads · 30 days
0
Access
Public
Updated Apr 24, 2026
Repo size
46.7 MB
Likes
0
Public
Click a slice to open those files.
.safetensors35.2 MB · 76%
From the Hugging Face model README
A conversational LLM fine-tuned for Kashmiri language in Nastaliq script (نَستَعلیٖق).
| Base Model | Qwen/Qwen2.5-0.5B-Instruct |
| Method | Two-stage LoRA: Continued Pretraining + Instruction Fine-tuning |
| Language | Kashmiri (کٲشُر) - Nastaliq Script |
| Parameters | ~494M |
| License | Apache 2.0 |
Stage 1 — Continued Pretraining: 1,500 high-quality Kashmiri Nastaliq texts from Wikipedia, language corpus, and text datasets.
Stage 2 — Instruction Fine-tuning: 34 hand-crafted Kashmiri Nastaliq Q&A pairs covering greetings, geography, culture, food, history, and general knowledge.
from peft import AutoPeftModelForCausalLM
from transformers import AutoTokenizer
import torch
model = AutoPeftModelForCausalLM.from_pretrained("megaketu555/kashmiri-nastaliq-llm")
tokenizer = AutoTokenizer.from_pretrained("megaketu555/kashmiri-nastaliq-llm")
messages = [{"role": "user", "content": "کٔشیٖر کیا چھُ؟"}]
text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(text, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=200, temperature=0.7, do_sample=True)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[1]:], skip_special_tokens=True))