WeMM-Video-9B

Model updates Hugging Face 模型 · tencent Source language: Chinese

WeMM-Video-9B is a dense 9B bilingual Chinese-English checkpoint for joint audio-visual understanding and reasoning, accepting text, images, audio, and video. The family was pretrained on about 1.2 trillion multimodal tokens and released under Apache 2.0 with BF16 weights plus FP8, AWQ, GGUF, and NVFP4 variants.

Full text

WeMM-Video-9B

Overview

WeMM-Video-9B is the dense 9B checkpoint in the WeMM-Video family. WeMM-Video is a Chinese–English bilingual family for joint audio-visual understanding and reasoning. It accepts text, images, audio, and video and generates text responses.

Weights · BF16 Safetensors · 19.54 GiB of weight files Quantized variants · FP8 Block128 · AWQ W4A16 · GGUF Q4_K_M · NVFP4 W4A16

Capabilities

WeMM-Video can combine video frames with the original soundtrack to describe events, transcribe speech, and reason about what is seen and heard. It also accepts silent video, standalone audio, images, and text, with responses in Chinese or English.

Architecture and training

Vision and audio encoders turn media into tokens that the language model processes alongside text. Video frames carry timestamps. The family was pretrained on approximately 1.2 trillion multimodal tokens, then post-trained with multi-teacher OPD, anti-repetition MrDPO, and specialist reinforcement learning.

See the project overview for family details.

Quick Start

Set up the Transformers environment. From the WeMM-Video repository root, run the example on an MP4 with a soundtrack. The script loads the model and processor from the Hub and processes both the frames and the original audio track.

Use --revision to pin the model and processor to a Hub commit. Size GPU memory for the full weight files, media inputs, and generation cache.

Input examples in Python

Run from the project root to use the bundled media. Each call sends one request. For video, video processes frames alone and video_audio includes the original soundtrack. Both take the file path in the video field.

Chat Modes

Instruct mode is the default. When calling processor.apply_chat_template, pass enable_thinking=False for a direct response or enable_thinking=True for a reasoning response.

Serving

Install the matching WeMM runtime patch. The serving guide lists the dependency profiles and Docker images. Choose one server below.

vLLM

Use patched vLLM 0.18.0 for the command below. The release guide also covers patched vLLM 0.31.0 for BF16. See the vLLM recipe for setup details.

SGLang

Use the WeMM BF16 recipe for patched SGLang 0.5.16 or 0.5.19.

Input examples for the serving API

With the server running on port 8000, use the bundled request client from the project root in a second terminal. It sends media inline to the OpenAI-compatible chat endpoint.

Input Request content Text Text string in messages[].content Image image_url with image_url.url Audio audio_url with audio_url.url Video frames video_url with video_url.use_audio set to false Video with soundtrack video_url with video_url.use_audio set to true

Set fps and max_frames inside video_url to control sampling for a clip. Set chat_template_kwargs.enable_thinking to true for Thinking mode. See the complete video request.

Validation

Patched vLLM loaded the checkpoint on H20 and completed generation checks for text, image, audio, and video with a soundtrack. The local checkpoint also passed direct Transformers loading and a CPU text generation check.

Evaluation

See the WeMM-Video-35B-A3B evaluation results for the benchmark comparison in the technical report.

License

Apache License 2.0. See the included license and the collection license and third-party attribution. Third-party components remain subject to their original terms.

Sources

← Back to Daily AI Report