vram.run Models Hardware Providers Cloud State of Inference
API provider data is live · Hardware & cloud pricing curated 2026-02-23

detr-resnet-101

by facebook

61M params · object-detection · 125 likes · 51.3k downloads

detr-resnet-101 is a 61M parameter model. At Q4 quantization it requires 0GB of VRAM. It runs comfortably on GeForce RTX 4090 (21639 tok/s), GeForce RTX 5090 (32451 tok/s), M4 Max 128GB (7918 tok/s).

Inference providers

Provider$/1M in$/1M outThroughput
HF Inference

GPU compatibility

GPUVRAMQ4 DecodeVerdict
GeForce RTX 409024GB21639 tok/scomfortable
GeForce RTX 509032GB32451 tok/scomfortable
M4 Max 128GB128GB7918 tok/scomfortable
M4 Pro 48GB48GB3959 tok/scomfortable
M4 Pro 24GB24GB3959 tok/scomfortable
A100 PCIe 80 GB80GB39647 tok/scomfortable
H100 SXM5 80 GB80GB79742 tok/scomfortable
GeForce RTX 309024GB19132 tok/scomfortable
Radeon RX 7900 XTX24GB15821 tok/scomfortable
GeForce RTX 408016GB15357 tok/scomfortable
Install CLI [email protected] Raw data · MIT · API data: live · HW/Cloud data: curated 2026-02-23 · v0.6.0