Text Generation
Transformers
TensorBoard
ONNX
Safetensors
English
llama
alignment-handbook
trl
sft
conversational
text-generation-inference
Instructions to use HuggingFaceTB/smollm-360M-instruct-add-basics with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use HuggingFaceTB/smollm-360M-instruct-add-basics with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="HuggingFaceTB/smollm-360M-instruct-add-basics") messages = [ {"role": "user", "content": "Who are you?"}, ] pipe(messages)# Load model directly from transformers import AutoTokenizer, AutoModelForMultimodalLM tokenizer = AutoTokenizer.from_pretrained("HuggingFaceTB/smollm-360M-instruct-add-basics") model = AutoModelForMultimodalLM.from_pretrained("HuggingFaceTB/smollm-360M-instruct-add-basics") 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 HuggingFaceTB/smollm-360M-instruct-add-basics with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "HuggingFaceTB/smollm-360M-instruct-add-basics" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "HuggingFaceTB/smollm-360M-instruct-add-basics", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/HuggingFaceTB/smollm-360M-instruct-add-basics
- SGLang
How to use HuggingFaceTB/smollm-360M-instruct-add-basics 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 "HuggingFaceTB/smollm-360M-instruct-add-basics" \ --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": "HuggingFaceTB/smollm-360M-instruct-add-basics", "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 "HuggingFaceTB/smollm-360M-instruct-add-basics" \ --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": "HuggingFaceTB/smollm-360M-instruct-add-basics", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use HuggingFaceTB/smollm-360M-instruct-add-basics with Docker Model Runner:
docker model run hf.co/HuggingFaceTB/smollm-360M-instruct-add-basics
File size: 3,064 Bytes
d26da8b d235430 d26da8b d235430 d26da8b d235430 d26da8b d235430 d26da8b d235430 d26da8b d235430 d26da8b d235430 d26da8b | 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 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 | from transformers import AutoModelForCausalLM, AutoTokenizer
BASE_PATH = "/fsx/loubna/projects/alignment-handbook/recipes/cosmo2/sft/data"
TEMPERATURE = 0.2
TOP_P = 0.9
CHECKPOINT = "HuggingFaceTB/smollm-350M-instruct-add-basics"
print(f"💾 Loading the model and tokenizer: {CHECKPOINT}...")
device = "cuda"
tokenizer = AutoTokenizer.from_pretrained(CHECKPOINT)
model_s = AutoModelForCausalLM.from_pretrained(CHECKPOINT).to(device)
print("🧪 Testing single-turn conversations...")
L = [
# Witing and general knowledge prompts
"Discuss the ethical implications of using AI in hiring processes.",
"Give me some tips to improve my time management skills?",
"Write a short dialogue between a customer and a waiter at a restaurant.",
"wassup?",
"Tell me a joke",
"Hi, what are some popular dishes from Japan?",
"What is the capital of Switzerland?",
"What is the capital of France?",
"What's the capital of Portugal?",
"What is the capital of Morocco?",
"How do I make pancakes?",
"Write a poem about Helium",
"Do you think it's important for a company to have a strong company culture? Why or why not?",
"What is your favorite book?",
"What is the most interesting fact you know?",
"What is your favorite movie?",
# Science prompts
"Can you tell me what is gravity?",
"Who discovered gravity?",
"How does a rainbow form?",
"What are the three states of matter?",
"Why is the sky blue?",
"What is the water cycle?",
"How do magnets work?",
"What is buoyancy?",
"What is the speed of light?",
"What's 2+2?",
"what's the sum of 2 and 2?",
"what's the sum of 2 and 3?",
"What is the term for the process by which plants make their own food?",
"If you have 8 apples and you give away 3, how many apples do you have left?",
# Python prompts
"How do I define a function in Python?",
"Can you explain what a dictionary is in Python?",
"Write a sort alrogithm in Python",
"Write a fibonacci sequence in Python",
"How do I read a file in Python?",
"How do I make everything uppercase in Python?",
"implement bubble sort in Python",
# Creative prompts
"Write a short story about a time traveler",
"Describe a futuristic city in three sentences",
"Describe a new color that doesn't exist",
"Create a slogan for a time machine company",
"Describe a world where plants can speak",
]
for i in range(len(L)):
print(f"🔮 {L[i]}")
messages = [{"role": "user", "content": L[i]}]
input_text = tokenizer.apply_chat_template(messages, tokenize=False)
inputs = tokenizer.encode(input_text, return_tensors="pt").to(device)
outputs = model_s.generate(
inputs, max_new_tokens=200, top_p=TOP_P, do_sample=True, temperature=TEMPERATURE
)
with open(
f"{BASE_PATH}/{CHECKPOINT.split('/')[-1]}_temp_{TEMPERATURE}_topp{TOP_P}.txt",
"a",
) as f:
f.write("=" * 50 + "\n")
f.write(tokenizer.decode(outputs[0]))
f.write("\n")
|