Instructions to use Riasok/kaana-if-opd with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Riasok/kaana-if-opd with Transformers:

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

pipe = pipeline("text-generation", model="Riasok/kaana-if-opd")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Riasok/kaana-if-opd")
model = AutoModelForCausalLM.from_pretrained("Riasok/kaana-if-opd", 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 Riasok/kaana-if-opd with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/Riasok/kaana-if-opd

SGLang

How to use Riasok/kaana-if-opd 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 "Riasok/kaana-if-opd" \
    --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": "Riasok/kaana-if-opd",
		"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 "Riasok/kaana-if-opd" \
        --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": "Riasok/kaana-if-opd",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Riasok/kaana-if-opd with Docker Model Runner:
```
docker model run hf.co/Riasok/kaana-if-opd
```

kaana-if-opd

File size: 1,559 Bytes

61a8b19

---
base_model: gyubinc/sft
library_name: transformers
pipeline_tag: text-generation
tags:
- generated_from_trainer
- trl
- minillm
- opd
- kanana
license: other
---

# kaana-if-opd

This is an open text-generation model fine-tuned from `gyubinc/sft` with OPD/TRL.

## Training

- Student/base model: `gyubinc/sft`
- Teacher model: `gyubinc/autoif`
- Dataset: `gyubinc/IF`
- Method: OPD / MiniLLM-style distillation (`distill_mode=tml`)
- Precision: bfloat16
- Training checkpoint: `checkpoint-50`
- Max planned steps: 100
- Saved global step: 50
- Effective batch size: 128
- Learning rate: 1e-6
- Max prompt length: 1024
- Max completion length: 2048

The source run was interrupted after step 62, and the latest complete saved Hugging Face checkpoint available locally was step 50. This repository contains only inference artifacts and excludes DeepSpeed optimizer state, RNG state, and trainer-resume files.

## Usage

```python
from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "riasok/kaana-if-opd"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype="auto",
    device_map="auto",
)

prompt = "Write a concise answer: What is instruction following?"
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
outputs = model.generate(**inputs, max_new_tokens=256)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
```

## Framework Versions

- TRL: 0.26.0
- Transformers: 4.57.1
- PyTorch: 2.8.0
- Datasets: 3.6.0
- Tokenizers: 0.22.2