You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

Resilient AI Challenge Round 1 Submission

Base model: mistralai/Voxtral-Mini-4B-Realtime-2602
Category: audio-to-text
Inference engine: vLLM
License: Apache-2.0, same as the original model

Methodology

This Round 1 submission prioritizes compatibility with the official evaluation harness.

The submitted model uses Voxtral Mini 4B Realtime as the primary model for inference. No architectural changes were made. No fine-tuning was performed after compression.

Applied optimization:

vLLM serving configuration
Reduced max_model_len to 20000 to reduce memory allocation during evaluation

Used Python 3.10 in a conda environment to test. Libraries used are in requirements.txt

Inference

The model is intended to be launched with:

vllm serve --config vllm_config.yaml

Downloads last month: -

Safetensors

Model size

4B params

Tensor type

BF16

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support