michael david PRO
clopeux
AI & ML interests
Gaussian Splat is where it's at
Recent Activity
upvoted a paper about 16 hours ago
Reducing the Offline-Streaming Gap for Unified ASR Transducer with Consistency Regularization reacted to nwaughachukwuma's post with 👍 about 16 hours ago
Can a text-only model + a vision toolkit (mm-ctx) match a native vision model?
We benchmarked 4 setups on 23 multimodal tasks (image, video, audio, PDF):
• gemini-3.5-flash (vision): 85.5
• glm-5.2 + mm-ctx (text-only): 75.9
• deepseek-v4-pro + mm-ctx (text-only): 68.8
• qwen3.6-35b-a3b (vision): 46.0
The best text-only setup closes 89% of the gap to the top vision model.
Full report: https://huggingface.co/blog/vlm-run/text-only-models-with-mm liked a model about 17 hours ago
nvidia/personaplex-7b-v1Organizations
None yet