SAM ViT Base โ€” BonfyreFPQ v12

Native .fpq v12 quantized weights for facebook/sam-vit-base.

Model Info

Property Value
Architecture Image Segmentation
Parameters 94M
Original Size 357 MB (safetensors)
FPQ v12 Size 84 MB
Compression 4.3ร—
Avg Cosine 0.999497
Worst Cosine 0.992606

Note: 8 rel_pos tensors (27ร—64) have reduced quality; all other tensors excellent

Format

BonfyreFPQ v12 native format: rANS entropy-coded E8 lattice coordinates with 6-bit packed tiles and FP16 scales.

Decode to safetensors

bonfyre-fpqx decode sam-vit-base-v12-fpq.fpq output.safetensors

Source

Converted from facebook/sam-vit-base using BonfyreFPQ v12.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support