GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
tokhey
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Run this model inference with full control and performance in your environment.
Model Details
Model Provider
tokhey
Model Tree
mistralai/Mistral-7B-Instruct-v0.3
Input Modalities
Output Modalities
Supported Functionality
The model is intended for generating English MCQs for educational purposes. Typical tasks include:
text
Generate 5 official Egyptian Ministry of Education English multiple-choice questions.Unit: 10Topic:At the AirportGrammar Focus:Past Perfect / Past Perfect ContinuousReturn only valid JSON.
json
{"questions": [{"statement": "...","correct_answer": "...","plausible_distractors": ["...","...","..."],"explanation": "..."}]}
The model was fine-tuned on a synthetic instruction dataset containing English examination questions inspired by the Egyptian Ministry of Education curriculum. The dataset includes:
All samples are formatted as multi-turn chat conversations.
python
from transformers import AutoTokenizer, AutoModelForCausalLMmodel_id = "tokhey/egyptian-mcq-generator-mistral-7b-synthetic"tokenizer = AutoTokenizer.from_pretrained(model_id)model = AutoModelForCausalLM.from_pretrained(model_id,torch_dtype="auto",device_map="auto",)prompt = "Generate 5 grammar questions about Past Perfect."inputs = tokenizer(prompt, return_tensors="pt").to(model.device)outputs = model.generate(**inputs,max_new_tokens=512,)print(tokenizer.decode(outputs[0], skip_special_tokens=True))
Although the model has been fine-tuned for educational assessment generation, it may occasionally:
The model has been qualitatively evaluated by comparing its outputs with the original Mistral-7B-Instruct-v0.3 base model. Fine-tuning improves:
Ahmed Eltokhy GitHub: https://github.com/ahmdeltoky03
LinkedIn: https://www.linkedin.com/in/ahmed-eltokhy-7b7a073ba