File size: 1,759 Bytes
516f4fc 633c18c e946c74 516f4fc 633c18c e946c74 633c18c 516f4fc 633c18c e946c74 516f4fc 633c18c e946c74 633c18c e6be6d1 99bf20e e6be6d1 99bf20e e6be6d1 99bf20e e6be6d1 633c18c 99bf20e 633c18c 99bf20e |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 |
### Finetuned Academic Question-Answering Model for ICSE Physics (Class 9 & 10)
This specialized large language model (LLM) is finetuned to provide precise and accurate answers to ICSE Physics questions for Classes 9 and 10. It is designed to assist students, educators, and content creators in understanding and exploring fundamental physics concepts aligned with the ICSE curriculum.
## Key Features
# 📚 Curriculum-Specific Training
Focused exclusively on ICSE Class 9 and 10 Physics topics, such as:
Motion
Work, Energy, and Power
Heat and Thermodynamics
Electricity and Magnetism
Light (Reflection and Refraction)
Sound
Modern Physics
# 🎯 Accurate and Concise Answers
Trained to deliver curriculum-aligned, student-friendly responses.
# Contextual Understanding
Handles specific and multi-part questions effectively, ensuring relevance and precision.
Example Usage
python
Copy code
from transformers import pipeline
# Load the model from Hugging Face
```python
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained(
"pitangent-ds/academic_phy",
load_in_4bit=True, # Quantized model
device_map="auto",
# llm_int8_enable_fp32_cpu_offload=True
)
tokenizer = AutoTokenizer.from_pretrained("pitangent-ds/academic_phy")
```
# Perform inference
```python
text = "What are units ?"
inputs = tokenizer(text, return_tensors="pt")
outputs = model.generate(**inputs)
decoded_output = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(decoded_output)
```
# Training Details
Dataset: Curated ICSE Physics content for Classes 9 and 10 textbooks
Loss Function: Cross-entropy loss
Final Training Loss: 0.88
Training Framework: PyTorch, Hugging Face Transformers
|