Instructions to use Kaito-F/qwen3-4b-agentbench-opd with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Kaito-F/qwen3-4b-agentbench-opd with Transformers:

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

pipe = pipeline("text-generation", model="Kaito-F/qwen3-4b-agentbench-opd")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Kaito-F/qwen3-4b-agentbench-opd")
model = AutoModelForCausalLM.from_pretrained("Kaito-F/qwen3-4b-agentbench-opd")
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 Kaito-F/qwen3-4b-agentbench-opd with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Kaito-F/qwen3-4b-agentbench-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": "Kaito-F/qwen3-4b-agentbench-opd",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Kaito-F/qwen3-4b-agentbench-opd

SGLang

How to use Kaito-F/qwen3-4b-agentbench-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 "Kaito-F/qwen3-4b-agentbench-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": "Kaito-F/qwen3-4b-agentbench-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 "Kaito-F/qwen3-4b-agentbench-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": "Kaito-F/qwen3-4b-agentbench-opd",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Kaito-F/qwen3-4b-agentbench-opd with Docker Model Runner:
```
docker model run hf.co/Kaito-F/qwen3-4b-agentbench-opd
```

qwen3-4b-agentbench-opd-merged

This repository provides a merged model (LoRA weights merged into base) trained from Kaito-F/qwen3-4b-instruct-lora-v2 using OPD (On-Policy Distillation).

This is a standalone model - no adapter loading required.

Training Pipeline

Stage 1: SFT - Supervised fine-tuning on agent trajectory data
Stage 2: OPD - On-Policy Distillation from teacher model

Training Objective

This model is trained to improve multi-turn agent task performance on ALFWorld (household tasks) and DBBench (database operations) through knowledge distillation from a larger teacher model.

The OPD method reduces exposure bias by training on the student's own generations while guided by teacher model probabilities.

Training Configuration

Stage 2: OPD

Student base: Kaito-F/qwen3-4b-instruct-lora-v2
Teacher model: Qwen/Qwen3-30B-A3B-Instruct-2507
Method: On-Policy Distillation with LoRA
Max sequence length: 2048
Steps: 100
Learning rate: 1e-05
LoRA: r=32, alpha=64

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

model_id = "Kaito-F/qwen3-4b-agentbench-opd"

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype=torch.float16,
    device_map="auto",
)
# No PeftModel needed - weights are already merged!

Sources & Terms (IMPORTANT)

Training data:

u-10bei/sft_alfworld_trajectory_dataset_v4
u-10bei/dbbench_sft_dataset_react_v4

Dataset License: MIT License. This dataset is used and distributed under the terms of the MIT License. Compliance: Users must comply with the MIT license (including copyright notice) and the base model's original terms of use.

Downloads last month: 1

Safetensors

Model size

4B params

Tensor type

F32

BF16

Model tree for Kaito-F/qwen3-4b-agentbench-opd

Base model

Qwen/Qwen3-4B-Instruct-2507

Adapter

Kaito-F/qwen3-4b-instruct-lora-v2

Quantized

(1)

this model

Kaito-F
/

qwen3-4b-agentbench-opd