Instructions to use yingfanbot/gsm-lotus-llama3b-codi with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use yingfanbot/gsm-lotus-llama3b-codi with Transformers:

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

pipe = pipeline("text-generation", model="yingfanbot/gsm-lotus-llama3b-codi")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("yingfanbot/gsm-lotus-llama3b-codi")
model = AutoModelForCausalLM.from_pretrained("yingfanbot/gsm-lotus-llama3b-codi")
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 yingfanbot/gsm-lotus-llama3b-codi with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "yingfanbot/gsm-lotus-llama3b-codi"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "yingfanbot/gsm-lotus-llama3b-codi",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/yingfanbot/gsm-lotus-llama3b-codi

SGLang

How to use yingfanbot/gsm-lotus-llama3b-codi 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 "yingfanbot/gsm-lotus-llama3b-codi" \
    --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": "yingfanbot/gsm-lotus-llama3b-codi",
		"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 "yingfanbot/gsm-lotus-llama3b-codi" \
        --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": "yingfanbot/gsm-lotus-llama3b-codi",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use yingfanbot/gsm-lotus-llama3b-codi with Docker Model Runner:
```
docker model run hf.co/yingfanbot/gsm-lotus-llama3b-codi
```

gsm-lotus-llama3b-codi

LOTUS + CODI (adds CODI trajectory distillation) checkpoint fine-tuned from meta-llama/Llama-3.2-3B-Instruct on GSM8k-Aug, from the paper Bridging the Gap Between Latent and Explicit Reasoning with Looped Transformers. This variant was trained with an extra CODI's trajectory-level distillation loss.

GSM8K (GSM8k-Aug) test accuracy: 70.58% (931/1319)
Latent config: K = 6 blocks, c_thought = 25 tokens/block, n_looped_iters = 6
Base: meta-llama/Llama-3.2-3B-Instruct (vocab 128256 -> 128259 for 3 latent tokens)

Loading

from_pretrained loads the weights only — the looped padded architecture needs the LOTUS code (code repo).

Reproduce the number above

Run in the pinned env (torch 2.7 / transformers 4.46.2). This loads the safetensors straight from this repo and runs the latent loop — no separate checkpoint file needed:

python scripts/eval.py \
  --model_id yingfanbot/gsm-lotus-llama3b-codi \
  --datasets gsm8k --n_looped_iters 6 --c_thought 25 --bf16

Yields 70.58% on GSM8k-Aug (verified by loading this repo's safetensors directly).

Citation

@article{fan2026bridging,
  title={Bridging the Gap Between Latent and Explicit Reasoning with Looped Transformers},
  author={Fan, Ying and Svete, Anej and Lee, Kangwook},
  journal={arXiv preprint arXiv:2606.31779},
  year={2026}
}