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 / README.md

Riasok

Add files using upload-large-folder tool

61a8b19 verified 4 months ago

preview code

Raw

History Blame Contribute Delete

1.56 kB

metadata

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

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