Downloads · 30 days
22
2% of all-time downloads
livesiyam/Qwen2.5-7B-Bangla-Physics-NCTB
Qwen2.5-7B-Bangla-Physics-NCTB is a text generation model from livesiyam. Use it when you need the model to write or continue text. It is set up for peft. The card lists the license as apache-2.0.
This is a fine-tuned version of Qwen-2.5-7B-Instruct designed to function as an AI Physics Tutor for Bangladeshi students. It is specifically trained on the National Curriculum and Textbook Board (NCTB) physics syllab…
Downloads · 30 days
22
2% of all-time downloads
All-time downloads
1.2K
Public
Repo size
20 GB
Likes
1
Public
Click a slice to open those files.
.gguf20 GB · 100%
From the Hugging Face model README
This is a fine-tuned version of Qwen-2.5-7B-Instruct designed to function as an AI Physics Tutor for Bangladeshi students. It is specifically trained on the National Curriculum and Textbook Board (NCTB) physics syllabus for Class 9 and 10. The model explains complex physics concepts and solves curriculum-based problems in clear, academic Bangla.
The model is optimized for educational use cases in Bangladesh, including:
This model follows the ChatML prompt structure. To ensure the model stays in its "Teacher" persona, use the following system prompt:
System Prompt:
তুমি একজন অভিজ্ঞ পদার্থবিজ্ঞান শিক্ষক। উত্তরগুলো সহজ এবং স্পষ্ট বাংলায় দাও।
Example Prompt:
তড়িৎ চৌম্বক আবেশ কাকে বলে?
The model was fine-tuned on a high-quality dataset of Bangla Physics questions and answers.
| Parameter | Value |
|---|---|
| Optimizer | Paged AdamW 32-bit |
| Learning Rate | 1e-4 |
| Epochs | 3 |
| Batch Size | 1 (Grad Accumulation: 4) |
| Quantization | 4-bit NormalFloat (NF4) |
| LoRA Rank (r) | 16 |
| LoRA Alpha | 32 |
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
model_id = "livesiyam/Qwen2.5-7B-Bangla-Physics-NCTB"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
model_id,
torch_dtype=torch.float16,
device_map="auto"
)