Instructions to use hishab/titulm-llama-3.2-3b-v2.0 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use hishab/titulm-llama-3.2-3b-v2.0 with Transformers:

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

pipe = pipeline("text-generation", model="hishab/titulm-llama-3.2-3b-v2.0")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("hishab/titulm-llama-3.2-3b-v2.0")
model = AutoModelForCausalLM.from_pretrained("hishab/titulm-llama-3.2-3b-v2.0", 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]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use hishab/titulm-llama-3.2-3b-v2.0 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "hishab/titulm-llama-3.2-3b-v2.0"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "hishab/titulm-llama-3.2-3b-v2.0",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/hishab/titulm-llama-3.2-3b-v2.0

SGLang

How to use hishab/titulm-llama-3.2-3b-v2.0 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 "hishab/titulm-llama-3.2-3b-v2.0" \
    --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": "hishab/titulm-llama-3.2-3b-v2.0",
		"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 "hishab/titulm-llama-3.2-3b-v2.0" \
        --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": "hishab/titulm-llama-3.2-3b-v2.0",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use hishab/titulm-llama-3.2-3b-v2.0 with Docker Model Runner:
```
docker model run hf.co/hishab/titulm-llama-3.2-3b-v2.0
```

lm-evaluation-harness tasks?

by dipta007 - opened Mar 8, 2025

Discussion

dipta007

Mar 8, 2025

In the paper, it was mentioned that lm-evaluation-harness was used to evaluate.

Did you guys add the task in the lm-evaluation-harness repo or do you have any fork of the repo, so that we can use the same (prompt, scoring) in our paper for fair comparison?

Thanks

sagorsarker

Mar 8, 2025

Yes, here is the fork repo that contains the task.
https://github.com/hishab-nlp/lm-evaluation-harness/tree/main

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment