Instructions to use PleIAs/Baguettotron with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use PleIAs/Baguettotron with Transformers:

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

pipe = pipeline("text-generation", model="PleIAs/Baguettotron")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("PleIAs/Baguettotron")
model = AutoModelForCausalLM.from_pretrained("PleIAs/Baguettotron", 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 PleIAs/Baguettotron with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/PleIAs/Baguettotron

SGLang

How to use PleIAs/Baguettotron 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 "PleIAs/Baguettotron" \
    --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": "PleIAs/Baguettotron",
		"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 "PleIAs/Baguettotron" \
        --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": "PleIAs/Baguettotron",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use PleIAs/Baguettotron with Docker Model Runner:
```
docker model run hf.co/PleIAs/Baguettotron
```

Meaning of the reasoning tokens

by pdelobelle - opened Nov 16, 2025

Discussion

pdelobelle

PleIAs org Nov 16, 2025

Thanks for the nice model release! The special reasoning tokens seem really interesting. Finally something better than "Wait, ...".

I was looking into the use of these reasoning tokens and I think most are strait-forward, except the ⟨H≈X.X⟩ and ↺. So I wanted to check what the intended meaning was. Sadly this was undocumented in the readme.

This is what I figured out:

Token	Function	Example from traces
`→`	Logical implication or derivation step	"Concert halls grew larger → Original instruments couldn't compete"
`↺`	Iterative refinement?	"Back to specifics ↺:" ; "further undermines trust → ↺"
`※`	Critical insights or questions	"※ Key question: equitable tolling vs US immunity"
`?maybe?`	Speculation?	"Exception cases? ?maybe? Ion pairs in solution..."
`●`	High confidence: well-established facts, strong empirical evidence	"Hummingbird-Heliconia coevolution ● well-documented"
`◐`	Medium/partial confidence: incomplete knowledge, tentative claims	"Information completeness: ◐ Missing: specific statistical data"
`○`	Low confidence: weak knowledge, missing information, speculative claims	"○ Exact reasoning details: covered general framework"
`⚠`	Identifies biases, problematic assumptions, ambiguities	"Response assumes long jump. ⚠ Domain specificity issue."
`☐`	Unverified hypothesis or pending verification step	"☐ → ☑ 'Whorfianism of the third kind'" (before verification)
`☑`	Verified claim or completed verification step	"☐ → ☑ Strong theoretical grounding in linguistic relativity"
`✓`	Validation, confirmation	"Logical consistency check: ✓ Multiple independent factors converge"
`∴`	Therefore	"∴ Simple factual verification with supporting context."
`⟨H≈X.X⟩`	Hallucination? Or hypothesis strength? (0.1-1.8)	"⟨H≈0.8⟩ Island biogeography principles... empty niches..."

Pclanglais

PleIAs org Nov 17, 2025

Thanks for opening this. Readme now integrates a presentation of special reasoning tokens.

Pclanglais changed discussion status to closed Nov 17, 2025

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment