Instructions to use Taeckyung/TISD-OLMo3-7B-Instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Taeckyung/TISD-OLMo3-7B-Instruct with Transformers:

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

pipe = pipeline("text-generation", model="Taeckyung/TISD-OLMo3-7B-Instruct")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Taeckyung/TISD-OLMo3-7B-Instruct")
model = AutoModelForCausalLM.from_pretrained("Taeckyung/TISD-OLMo3-7B-Instruct", 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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Taeckyung/TISD-OLMo3-7B-Instruct with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Taeckyung/TISD-OLMo3-7B-Instruct"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Taeckyung/TISD-OLMo3-7B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Taeckyung/TISD-OLMo3-7B-Instruct

SGLang

How to use Taeckyung/TISD-OLMo3-7B-Instruct 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 "Taeckyung/TISD-OLMo3-7B-Instruct" \
    --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": "Taeckyung/TISD-OLMo3-7B-Instruct",
		"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 "Taeckyung/TISD-OLMo3-7B-Instruct" \
        --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": "Taeckyung/TISD-OLMo3-7B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Taeckyung/TISD-OLMo3-7B-Instruct with Docker Model Runner:
```
docker model run hf.co/Taeckyung/TISD-OLMo3-7B-Instruct
```

TISD-OLMo3-7B-Instruct

TL;DR: allenai/Olmo-3-7B-Instruct trained for 80 steps with TISD (Trajectory-Intervention Self-Distillation) on LiveCodeBench v6 with rich execution feedback.

Paper (arXiv:2609.30878) | Code | Project page

What is TISD?

On-policy self-distillation trains a student toward a teacher that is the same model conditioned on privileged context (a successful sibling rollout or execution feedback). TISD changes which trajectories are distilled: for each failed rollout it

finds the position of peak teacher-student divergence and forces the teacher's preferred token there,
lets the student regenerate the rest of the response, and
distills the resulting trajectory toward the privileged-context teacher with the SDPO objective.

Results (LiveCodeBench v6, 131 problems)

Model	Avg@4 (%)	Pass@4 (%)
`allenai/Olmo-3-7B-Instruct` (base)	32.1	42.7
TISD-OLMo3-7B-Instruct (this checkpoint, step 80)	51.0	53.4

Avg@4 is the mean test-case pass rate over 4 samples per problem; Pass@4 counts a problem as solved if any of the 4 samples passes all hidden tests. Numbers are for this exact checkpoint (training step 80). Paper Table 1 reports the best validation step, averaged over runs where several exist.

Training setup

Item	Value
Data	LiveCodeBench v6 (train split exposes 50% of each problem's tests)
Steps	80 (batch of 32 prompts, 8 rollouts each)
Learning rate	1e-6
Hardware	NVIDIA B200 GPUs, verl FSDP
Weights	bf16, merged from the FSDP actor checkpoint

Full hyperparameters are in the paper appendix and the code repository.

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

repo = "Taeckyung/TISD-OLMo3-7B-Instruct"
tokenizer = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(repo, torch_dtype="bfloat16", device_map="auto")

The model uses the base model's chat template. Loading requires transformers>=4.57 (OLMo 3 architecture).

License

Apache 2.0, the same as the base model allenai/Olmo-3-7B-Instruct. This is a modified (fine-tuned) version of the base model.

Citation

@misc{lee2026tisd,
  title  = {{TISD}: On-Policy Self-Distillation with Trajectory Intervention},
  author = {Lee, Taeckyung and Amankos, Rinat and Kim, Jeonghye and Yoon, Hyungjun and Jin, Woogyeol and Lee, Sung-Ju},
  year   = {2026},
  eprint = {2609.30878},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/2609.30878}
}