Instructions to use umarigan/llama3.2-1B-fin with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use umarigan/llama3.2-1B-fin with Transformers:

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

pipe = pipeline("text-generation", model="umarigan/llama3.2-1B-fin")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("umarigan/llama3.2-1B-fin")
model = AutoModelForCausalLM.from_pretrained("umarigan/llama3.2-1B-fin")
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

vLLM

How to use umarigan/llama3.2-1B-fin with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "umarigan/llama3.2-1B-fin"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "umarigan/llama3.2-1B-fin",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/umarigan/llama3.2-1B-fin

SGLang

How to use umarigan/llama3.2-1B-fin 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 "umarigan/llama3.2-1B-fin" \
    --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": "umarigan/llama3.2-1B-fin",
		"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 "umarigan/llama3.2-1B-fin" \
        --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": "umarigan/llama3.2-1B-fin",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Unsloth Studio new

How to use umarigan/llama3.2-1B-fin with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for umarigan/llama3.2-1B-fin to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for umarigan/llama3.2-1B-fin to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for umarigan/llama3.2-1B-fin to start chatting

Load model with FastModel

pip install unsloth
from unsloth import FastModel
model, tokenizer = FastModel.from_pretrained(
    model_name="umarigan/llama3.2-1B-fin",
    max_seq_length=2048,
)

Docker Model Runner
How to use umarigan/llama3.2-1B-fin with Docker Model Runner:
```
docker model run hf.co/umarigan/llama3.2-1B-fin
```

Small model for finance use cases fine-tuned on a QA dataset. Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.

Load model directly

from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("umarigan/llama3.2-1B-fin")
model = AutoModelForCausalLM.from_pretrained("umarigan/llama3.2-1B-fin")

alpaca_prompt = """Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.
### Instruction:
{}
### Input:
{}
### Response:
{}"""
inputs = tokenizer(
[
    alpaca_prompt.format(
        "How this article will effect stock price of Pfizer", # Görev
        """(Reuters) -Drugmaker Pfizer on Tuesday said it was expecting its 2025 profits to be nearly in line with Wall Street expectations as it seeks to fend off criticism over its turnaround strategy from activist hedge fund Starboard Value.
          Shares of the company rose nearly 3% in premarket trading.
          The company expects adjusted profit of $2.80 to $3 per share, compared with analysts' average estimate of $2.88 per share, according to data compiled by LSEG.
          Pfizer has been reining in costs and cutting debt by shedding non-core businesses as its rebuilds itself post a sharp slump in sales of COVID-19 products.
          Its shares have fallen nearly 12% this year and trade at about half their value during the peak of the COVID-19 pandemic as revenue slows down and several top-selling drugs come off patent.
          That has left it open to investor criticism, with Starboard in October saying that Pfizer's management has over-spent on big acquisitions and failed to produce profitable new drugs from those deals or from its internal research and development.
          Pfizer forecast 2025 revenue in the range of $61 billion to $64 billion, compared with the estimates of $63.26 billion.
          The company is slated to hold a conference call with analysts later in the day to discuss the forecast.
          (Reporting by Mariam Sunny and Manas Mishra; Editing by Anil D'Silva)
        """, # İnput
        "", # Response - Leave empty!
    )
], return_tensors = "pt")
outputs = model.generate(**inputs, max_new_tokens = 64, use_cache = True)
tokenizer.batch_decode(outputs)
Output:
### Response:\nPfizer is expected to see a significant increase in its profits in 2025.
The company is planning to hold a conference call with analysts later in the day to discuss the forecast.<|eot_id|>

Downloads last month: 1

umarigan
/

llama3.2-1B-fin

Load model directly

Dataset used to train umarigan/llama3.2-1B-fin