Skip to content

SmolVLM2-256M-Video-Instruct-mlx

mlx-community/SmolVLM2-256M-Video-Instruct-mlx
Video + text · transformers
Open on Hugging Face
Parameters
256M
Weights (est.)
≈ 513 MB
Mac memory (est.)
8 GB+

Size is estimated from the parameter count and quantisation. Memory allows a fifth extra for context plus 2 GB for the system; long contexts need more.

Details

Base model
HuggingFaceTB/SmolLM2-360M-Instruct (finetune)
Vendor
Hugging Face
Task
Video + text
Quantisation
Not stated
Licence
apache-2.0
Downloads
717 in the last 30 days · 13K all time · 10 likes
Published
2025-02-17 · updated 2025-02-26 (1y ago)
Languages
en
Tags
smolvlm, image-text-to-text

Other MLX builds of HuggingFaceTB/SmolLM2-360M-Instruct

Build Quantisation Weights Mac memory Downloads
SmolVLM-500M-Instruct-3bit 3-bit ≈ 223 MB 8 GB+ 47
SmolVLM-500M-Instruct-4bit 4-bit ≈ 287 MB 8 GB+ 385
SmolLM2-360M-Instruct-6bit 6-bit ≈ 294 MB 8 GB+ 103
SmolVLM-500M-Instruct-6bit 6-bit ≈ 413 MB 8 GB+ 44
SmolVLM2-256M-Video-Instruct-mlx (this one) — ≈ 513 MB 8 GB+ 717
SmolVLM-500M-Instruct-8bit 8-bit ≈ 540 MB 8 GB+ 56
SmolVLM2-500M-Video-Instruct-mlx-8bit-skip-vision 8-bit ≈ 620 MB 8 GB+ 238
SmolLM2-360M-Instruct — ≈ 724 MB 8 GB+ 681
SmolLM2-360M-Instruct-bf16-mlx bf16 ≈ 724 MB 8 GB+ 63
SmolVLM2-500M-Video-Instruct-mlx — ≈ 1.0 GB 8 GB+ 5.3K
SmolVLM-500M-Instruct-bf16 bf16 ≈ 1.0 GB 8 GB+ 109