Add the Qwen3.5-9B demo video

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LrUG2MPyVgRWHb7y6iKMf4

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+assets/demo.mp4 filter=lfs diff=lfs merge=lfs -text

README.md

@@ -27,6 +27,14 @@ MMFlash draft model for speculative decoding of [`Qwen/Qwen3.8-27B`](https://hug
 
 MMFlash builds on [DFlash](https://github.com/z-lab/dflash) block-diffusion drafting: the draft reads hidden states from five layers of the target and proposes a block of 16 tokens in one forward pass, and the target verifies the block in one forward pass. MMFlash keeps the DFlash architecture and is trained on multimodal data with a loss weighted by per-token visual-grounding (visual-KL) scores.
 
+<div align="center">
+<video autoplay muted loop playsinline controls width="100%">
+  <source src="https://huggingface.co/FSCCS/Qwen3.8-27B-MMFlash/resolve/main/assets/demo.mp4" type="video/mp4">
+</video>
+</div>
+
+<p align="center"><sub>Qwen3.5-9B on a MathVista question, recorded in real time on one NVIDIA H200 with SGLang (greedy decoding).</sub></p>
+
 > [!TIP]
 > Need a drafter for another model? Email us at [fscnkucs@gmail.com](mailto:fscnkucs@gmail.com) and we can help train one. We will keep updating these releases, and a Qwen4 checkpoint is planned.
 

assets/demo.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ab4f4826ed2dd5e33335ff49c77ed5860f56db1e4479ed5db4c559ba2c32ddf2
+size 2234619