Instructions to use Retreatcost/Voxtral-TCR1-4b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Retreatcost/Voxtral-TCR1-4b with Transformers:

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

pipe = pipeline("text-generation", model="Retreatcost/Voxtral-TCR1-4b")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Retreatcost/Voxtral-TCR1-4b")
model = AutoModelForCausalLM.from_pretrained("Retreatcost/Voxtral-TCR1-4b")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Retreatcost/Voxtral-TCR1-4b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Retreatcost/Voxtral-TCR1-4b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Retreatcost/Voxtral-TCR1-4b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Retreatcost/Voxtral-TCR1-4b

SGLang

How to use Retreatcost/Voxtral-TCR1-4b 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 "Retreatcost/Voxtral-TCR1-4b" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Retreatcost/Voxtral-TCR1-4b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "Retreatcost/Voxtral-TCR1-4b" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Retreatcost/Voxtral-TCR1-4b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Retreatcost/Voxtral-TCR1-4b with Docker Model Runner:
```
docker model run hf.co/Retreatcost/Voxtral-TCR1-4b
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Voxtral-TCR1-4b

This is a Voxtral model with removed voice module and finetuned to give it a custom task-concept reasoning pattern:

Question: <question>

Sub-tasks:
    1.
    2.
    3.

Key concepts:
    -
    -
    -

Use ChatML formatting, force thinking mode in your favourite front-end (prefill with <think> token).

Temps in range 0.6-0.8 seem to work reasonably well.

This is an experiment to see if thinking/reasoning could be bootstrapped from 0 without any reasoning datasets whatsoever. The answer is yes.

This model is trained on purely artificial data of non-reasoning models.

Downloads last month: 1

Safetensors

Model size

4B params

Tensor type

BF16

Model tree for Retreatcost/Voxtral-TCR1-4b

Base model

mistralai/Voxtral-Mini-3B-2507

Finetuned

Leoxxxxh/Voxtral-Mini-3B-2507-TextOnly

Finetuned

(1)

this model

Quantizations

3 models

Collection including Retreatcost/Voxtral-TCR1-4b

Mad science

Collection

Experimental and probably broken in some ways • 9 items • Updated Nov 16, 2025 • 2