Instructions to use codercat/AtmosRegime-Qwen3-8B-Distill with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use codercat/AtmosRegime-Qwen3-8B-Distill with Transformers:

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

pipe = pipeline("text-generation", model="codercat/AtmosRegime-Qwen3-8B-Distill")
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("codercat/AtmosRegime-Qwen3-8B-Distill")
model = AutoModelForCausalLM.from_pretrained("codercat/AtmosRegime-Qwen3-8B-Distill", 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 codercat/AtmosRegime-Qwen3-8B-Distill with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "codercat/AtmosRegime-Qwen3-8B-Distill"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "codercat/AtmosRegime-Qwen3-8B-Distill",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/codercat/AtmosRegime-Qwen3-8B-Distill

SGLang

How to use codercat/AtmosRegime-Qwen3-8B-Distill 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 "codercat/AtmosRegime-Qwen3-8B-Distill" \
    --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": "codercat/AtmosRegime-Qwen3-8B-Distill",
		"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 "codercat/AtmosRegime-Qwen3-8B-Distill" \
        --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": "codercat/AtmosRegime-Qwen3-8B-Distill",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use codercat/AtmosRegime-Qwen3-8B-Distill with Docker Model Runner:
```
docker model run hf.co/codercat/AtmosRegime-Qwen3-8B-Distill
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

AtmosRegime-Qwen3-8B-Distill

Qwen3-8B trained by on-policy distillation from a teacher given the correct formula, from the paper Language models misapply atmospheric formulas until trained on a few worked examples. In the paper, this route raises outside-regime accuracy from 0% to 71% on mechanisms the base model never solved but did solve once told the formula.

Code and data: https://github.com/acodercat/AtmosRegime. The other models from the paper are in the AtmosRegime collection.

Training

Student: Qwen3-8B, thinking off.
Teacher: a frozen copy of Qwen3-8B whose prompt also contains the correct formula. The student never sees the formula; on its own samples it is pulled towards the teacher's token distribution (forward KL over the teacher's 32 most probable tokens), alongside the answer reward.
Settings: verl, learning rate 1e-6, 128 problems × 8 samples per step, 120 steps, no dynamic sampling.
Data: training fold 1 of the study's templates (29 of the 70 mechanisms); seed 42.

Use

The model was trained on one prompt format and should be given the same one. It answers with a Python program that defines solve(). Run it with thinking disabled.

# git clone https://github.com/acodercat/AtmosRegime && cd AtmosRegime   (for the prompt)
from transformers import AutoModelForCausalLM, AutoTokenizer
from atmoscoder_rl.prompts import chat

name = "codercat/AtmosRegime-Qwen3-8B-Distill"
tok = AutoTokenizer.from_pretrained(name)
model = AutoModelForCausalLM.from_pretrained(name, torch_dtype="auto", device_map="auto")

problem = "Atmospheric CO2 rises from 280 ppm to 560 ppm. Calculate the radiative forcing in W m-2."
text = tok.apply_chat_template(chat(problem), tokenize=False, add_generation_prompt=True, enable_thinking=False)
inputs = tok(text, return_tensors="pt").to(model.device)
out = model.generate(**inputs, max_new_tokens=3072, do_sample=True, temperature=0.3)
print(tok.decode(out[0][inputs.input_ids.shape[1]:], skip_special_tokens=True))

Execution and grading of the returned program are in atmoscoder_rl/evaluation/ of the repository.

Limitations

This checkpoint is one fold and one seed. Numbers in the paper average both folds and two seeds, so this model alone will not reproduce them exactly.
Distillation helped only where the teacher could solve the problem once told the formula.
It is not a general-purpose assistant for atmospheric science.

Citation

@unpublished{ran2026regime,
  title  = {Language models misapply atmospheric formulas until trained on a few worked examples},
  author = {Ran, Maohao and Zhang, Yanting and Ma, Chendong and Song, Jun},
  year   = {2026},
  note   = {Manuscript under review}
}