Instructions to use ISLAM-PO/MasryGPT_chat_FINALLY with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ISLAM-PO/MasryGPT_chat_FINALLY with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="ISLAM-PO/MasryGPT_chat_FINALLY")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("ISLAM-PO/MasryGPT_chat_FINALLY")
model = AutoModelForCausalLM.from_pretrained("ISLAM-PO/MasryGPT_chat_FINALLY", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

PEFT
How to use ISLAM-PO/MasryGPT_chat_FINALLY with PEFT:
```
Task type is invalid.
```
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use ISLAM-PO/MasryGPT_chat_FINALLY with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ISLAM-PO/MasryGPT_chat_FINALLY"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ISLAM-PO/MasryGPT_chat_FINALLY",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ISLAM-PO/MasryGPT_chat_FINALLY

SGLang

How to use ISLAM-PO/MasryGPT_chat_FINALLY with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "ISLAM-PO/MasryGPT_chat_FINALLY" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ISLAM-PO/MasryGPT_chat_FINALLY",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "ISLAM-PO/MasryGPT_chat_FINALLY" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ISLAM-PO/MasryGPT_chat_FINALLY",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Unsloth Desktop
Docker Model Runner
How to use ISLAM-PO/MasryGPT_chat_FINALLY with Docker Model Runner:
```
docker model run hf.co/ISLAM-PO/MasryGPT_chat_FINALLY
```

MasryGPT_chat_FINALLY / inference_example.py

ISLAM-PO

Upload 13 files

4e5fa10 verified 1 day ago

Raw

History Blame Contribute Delete

3.62 kB

	"""
	MasryGPT_chat Inference Example — English + مصري Bilingual
	مثال تشغيل MasryGPT_chat — انجليزي ومصري
	Model: ISLAM-PO/MasryGPT_chat (Qwen2.5-1.5B Egyptian Dialect)
	Path: C:\Users\RDP\Downloads\MasryGPT-merged\MasryGPT-merged (1)\kaggle\working\MasryGPT-merged
	"""

	# English: Install requirements
	# مصري: سطب المكتبات المطلوبة
	# pip install -r requirements.txt # or: pip install unsloth transformers accelerate bitsandbytes

	# ========== Option 1: Unsloth (Fastest, Recommended) / أسرع طريقة ==========
	from unsloth import FastLanguageModel

	# English: Load merged 16-bit model in 4-bit for fast inference on T4
	# مصري: حمّل الموديل المدمج 16-bit في وضع 4-bit عشان يشتغل بسرعة على T4
	model, tokenizer = FastLanguageModel.from_pretrained(
	model_name = "ISLAM-PO/MasryGPT_chat", # or local: "C:/.../MasryGPT-merged"
	# For local: model_name = r"C:\Users\RDP\Downloads\MasryGPT-merged\MasryGPT-merged (1)\kaggle\working\MasryGPT-merged"
	max_seq_length = 2048,
	dtype = None,
	load_in_4bit = True, # False = 9-11GB VRAM, True = 2-4GB VRAM
	)
	FastLanguageModel.for_inference(model) # 2x faster inference

	# English: Chat template - system prompt forces Egyptian
	# مصري: قالب الشات - رسالة النظام بتجبره يتكلم مصري
	messages = [
	{"role": "system", "content": "انت مساعد مصري تتكلم عامية مصرية فقط، ردك قصير وطبيعي زي ولاد البلد"},
	{"role": "user", "content": "ازيك يا باشا؟ احكيلي نكتة مصرية قصيرة"},
	# Try also: "يعني ايه عيش وملح بين الصحاب؟" / "اشرح لخواجة يعني ايه يخرم بيتك"
	]

	# English: Apply chat template with attention mask to fix warning
	# مصري: طبق قالب الشات مع attention_mask عشان تحذير Unsloth يختفي
	inputs = tokenizer.apply_chat_template(
	messages,
	tokenize=True,
	add_generation_prompt=True,
	return_tensors="pt",
	return_dict=True, # important to get attention_mask
	).to("cuda")

	# English: Generate - temperature 0.8 for creativity, 0.3 for facts
	# مصري: التوليد - 0.8 للهزار والإبداع، 0.3 للمعلومات الجادة
	outputs = model.generate(
	**inputs,
	max_new_tokens=256, # English: increase for longer jokes / مصري: زوده للنكت الطويلة
	temperature=0.8,
	top_p=0.95,
	do_sample=True,
	repetition_penalty=1.1, # reduces repetition / يقلل التكرار
	use_cache=True,
	)

	response = tokenizer.batch_decode(outputs[:, inputs["input_ids"].shape[1]:], skip_special_tokens=True)[0]
	print("="*40)
	print("MasryGPT Response / رد MasryGPT:")
	print(response)
	print("="*40)


	# ========== Option 2: Transformers (Standard) / الطريقة العادية ==========
	"""
	from transformers import AutoModelForCausalLM, AutoTokenizer

	model = AutoModelForCausalLM.from_pretrained(
	"ISLAM-PO/MasryGPT_chat",
	device_map="auto",
	trust_remote_code=True,
	)
	tokenizer = AutoTokenizer.from_pretrained("ISLAM-PO/MasryGPT_chat", fix_mistral_regex=True)
	# same generate code as above
	"""

	# ========== Quick Test Prompts / أسئلة سريعة للتجربة ==========
	test_prompts = [
	"ازيك عامل ايه؟",
	"احكيلي نكتة مصرية",
	"يعني ايه عيش وملح؟",
	"اشرح يعني ايه 'على راسي' بالمصري",
	"اعملي بوست فيسبوك مصري عن رمضان",
	]
	# Loop them with same messages logic