---
title: VLAC2-Qwen3VL-30B-A3B-Progress
canonical_url: "https://www.modelscope.ai/models/InternRobotics/VLAC2-Qwen3VL-30B-A3B-Progress"
md_url: "https://www.modelscope.ai/models/InternRobotics/VLAC2-Qwen3VL-30B-A3B-Progress.md"
repository: InternRobotics/VLAC2-Qwen3VL-30B-A3B-Progress
last_updated: 2026-06-23
pipeline_tag: image-text-to-text
tasks:
  - image-text-to-text
model_type:
  - qwen3_vl_moe
architectures:
  - Qwen3VLMoeForConditionalGeneration
base_model:
  - Qwen/Qwen3-VL-30B-A3B-Instruct
base_model_relation: finetune
parameters: 31.1B
tensor_type:
  - BF16
library_name:
  - transformer
  - safetensors
  - pytorch
frameworks:
  - pytorch
downloads: 141
stars: 0
tags:
  - robotics
  - video-understanding
  - multimodal
  - progress-estimation
---

# VLAC2-Qwen3VL-30B-A3B-Progress

> VLAC2-Qwen3VL-30B-A3B-Progress - An open-source model by InternRobotics on ModelScope. VLAC2-Qwen3VL-30B-A3B-Progress

InternRobotics/VLAC2-Qwen3VL-30B-A3B-Progress is a 31.1B-parameter image-text-to-text model on ModelScope. derived from Qwen/Qwen3-VL-30B-A3B-Instruct.

- **Repository**: InternRobotics/VLAC2-Qwen3VL-30B-A3B-Progress
- **Tasks**: image-text-to-text
- **Parameters**: 31.1B
- **Base model**: Qwen/Qwen3-VL-30B-A3B-Instruct
- **Tags**: robotics, video-understanding, multimodal, progress-estimation
- **Downloads**: 141
- **Stars**: 0
- **Last updated**: 2026-06-23

Source: https://www.modelscope.ai/models/InternRobotics/VLAC2-Qwen3VL-30B-A3B-Progress

---

# VLAC2-Qwen3VL-30B-A3B-Progress

VLAC2-Qwen3VL-30B-A3B-Progress is a video-language progress estimation model for robotic manipulation. Given a task description and a video, it predicts time-progress keypoints that can be aligned into a dense progress curve.

This release contains the inference-ready model weights, tokenizer and processor files, three bundled demo episodes, and a minimal local inference example.

## Model

- Base model: `Qwen/Qwen3-VL-30B-A3B-Instruct`
- Architecture: `Qwen3VLMoeForConditionalGeneration`
- Recommended VLAC model type in local evaluation scripts: `qwen3_moe_vl`
- Bundled quick start uses the `chunk_all` prompt and samples the input video at `2 Hz`

## Highlights

- Progress estimation for long-horizon manipulation videos
- Rollback and regression recognition in non-expert trajectories
- Bundled demos for both expert and non-expert episodes

## Use with Transformers

```python
from transformers import AutoModelForImageTextToText, AutoProcessor

model_id = "InternRobotics/VLAC2-Qwen3VL-30B-A3B-Progress"

processor = AutoProcessor.from_pretrained(model_id)
model = AutoModelForImageTextToText.from_pretrained(
    model_id,
    torch_dtype="auto",
    device_map="auto",
)
```

## Quick Start

Run one of the bundled examples from the model release root:

```bash
python quick_start/run_example.py --example-id example_01
python quick_start/run_example.py --example-id example_02
python quick_start/run_example.py --example-id example_03
```

For local VLAC benchmark scripts, use:

```bash
MODEL_PATH=/path/to/VLAC2-Qwen3VL-30B-A3B-Progress
MODEL_TYPE=qwen3_moe_vl
```

More details are provided in `quick_start/README.md`.