Datasets:

cfahlgren1
/

simready-usd-web-viewers

asset stringclasses 6 values	source stringclasses 6 values	root_format stringclasses 2 values	package_mb float32 2.3 219	three_0_174_status stringclasses 1 value	three_0_174_load_s float32	three_0_174_memory_mb int32	three_0_174_data_mb float32	three_0_174_triangles int32	three_0_174 imagewidth (px)	three_0_186_status stringclasses 3 values	three_0_186_load_s float32 0.15 1.31 ⌀	three_0_186_memory_mb int32 232 279 ⌀	three_0_186_data_mb float32 0 10.8 ⌀	three_0_186_triangles int32 0 134k ⌀	three_0_186 imagewidth (px) 256 256 ⌀	three_0_186_crawler_status stringclasses 3 values	three_0_186_crawler_load_s float32 0.15 1.31 ⌀	three_0_186_crawler_memory_mb int32 232 378 ⌀	three_0_186_crawler_data_mb float32 0 10.8 ⌀	three_0_186_crawler_triangles int32 0 134k ⌀	three_0_186_crawler imagewidth (px) 256 256 ⌀	needle_status stringclasses 2 values	needle_load_s float32 0.49 13.9	needle_memory_mb int32 1.12k 4.94k	needle_data_mb float32 0 218	needle_triangles int32 0 612k	tinyusdz_status stringclasses 3 values	tinyusdz_load_s float32 1.11 1.34 ⌀	tinyusdz_memory_mb int32 292 446 ⌀	tinyusdz_data_mb float32 0 10.8 ⌀	tinyusdz_triangles int32 0 135k ⌀	tinyusdz imagewidth (px) 256 256 ⌀	cinevva_status stringclasses 2 values	cinevva_load_s float32 2.21 2.7	cinevva_memory_mb int32 231 327	cinevva_data_mb float32 0 23.7	cinevva_triangles int32 0 134k	glb_status stringclasses 1 value	glb_load_s float32 1.16 1.2	glb_memory_mb int32 231 265	glb_data_mb float32 0 15.1	glb_triangles int32 260 612k
LG laptop	https://huggingface.co/datasets/LGElectronics/simready-assets/tree/main/laptop_17z90ur	usdc (binary)	11	❌ zip error	null	null	null	null	Not supported with pagination yet	✅ renders	1.31	279	10.8	134,228		✅ renders	1.31	378	10.8	134,228		✅ renders	1.17	1,314	10.8	134,228	✅ renders	1.34	446	10.8	134,545		✅ renders	2.21	327	10.8	134,228	✅ renders	1.19	231	4.6	134,228
Robotiq 2F-85 gripper	https://huggingface.co/datasets/Robotiq-Official/simready-assets/tree/main/Robotiq_2F_85	usda (text)	5.1	❌ zip error	null	null	null	null	Not supported with pagination yet	⚠️ blank	0.15	232	0	0		✅ renders	1.17	248	4	74,108		✅ renders	1.19	1,259	4.1	85,986	⚠️ blank	1.11	342	0.1	0		⚠️ blank	2.7	241	4.9	0	✅ renders	1.17	231	2.8	85,986
Standard Bots Thor	https://huggingface.co/datasets/standardbots/simready-thor/tree/main/standardbots_thor	usdc (binary)	47.400002	❌ zip error	null	null	null	null	Not supported with pagination yet	⚠️ blank	0.15	232	0	0		⚠️ blank	0.15	232	0	0		✅ renders	1.23	1,568	45.900002	612,452	⚠️ blank	1.14	367	2.4	0		⚠️ blank	2.61	231	0	6	✅ renders	1.17	242	15.1	612,452
NVIDIA IV pole	https://huggingface.co/datasets/nvidia/simready-assets/tree/main/IV_pole_a01	usdc (binary)	218.899994	❌ zip error	null	null	null	null	Not supported with pagination yet	⚠️ blank	0.15	232	0	0		⚠️ blank	0.15	232	0	0		✅ renders	13.93	4,941	217.899994	458,714	❌ compose error	null	null	null	null	Not supported with pagination yet	⚠️ blank	2.67	290	23.700001	6	✅ renders	1.17	265	10.8	458,714
NVIDIA office chair	https://huggingface.co/datasets/nvidia/simready-assets/tree/main/chair_office_fabric_manager_a01	usdc (binary)	35.299999	❌ zip error	null	null	null	null	Not supported with pagination yet	⚠️ blank	0.15	232	0	0		⚠️ blank	0.15	232	0	0		✅ renders	2.07	1,801	33	91,432	❌ compose error	null	null	null	null	Not supported with pagination yet	⚠️ blank	2.67	231	4.8	6	✅ renders	1.2	231	2.3	91,432
imagine.io railing	https://huggingface.co/datasets/imagineio/PhysicalAI-SimReady-Assets/tree/main/assets/railings/railing_1	usdc (binary)	2.3	❌ zip error	null	null	null	null	Not supported with pagination yet	❌ crash	null	null	null	null	Not supported with pagination yet	❌ crash	null	null	null	null	Not supported with pagination yet	⚠️ blank	0.49	1,118	0	0	✅ renders	1.16	292	0	260		⚠️ blank	2.62	231	0	2	✅ renders	1.16	232	0	260

SimReady assets in browser USD viewers

Six real SimReady OpenUSD packages from the Hub, loaded in five browser USD libraries and a pre-converted GLB baseline. Each cell is what the library drew.

Asset	three.js 0.174	three.js 0.186	three.js 0.186 + crawler	Needle	tinyusdz	cinevva usdjs	GLB (pre-converted)
LG laptop `usdc` · 11 MB	❌ zip error	✅ renders _{1.3 s · 279 MB}	✅ renders _{1.3 s · 378 MB}	✅ renders _{1.2 s · 1.3 GB}	✅ renders _{1.3 s · 446 MB}	✅ renders _{2.2 s · 327 MB}	✅ renders _{1.2 s · 231 MB}
Robotiq 2F-85 gripper `usda` · 5 MB	❌ zip error	⚠️ blank	✅ renders _{1.2 s · 248 MB}	✅ renders _{1.2 s · 1.2 GB}	⚠️ blank	⚠️ blank	✅ renders _{1.2 s · 231 MB}
Standard Bots Thor `usdc` · 47 MB	❌ zip error	⚠️ blank	⚠️ blank	✅ renders _{1.2 s · 1.5 GB}	⚠️ blank	⚠️ blank	✅ renders _{1.2 s · 242 MB}
NVIDIA IV pole `usdc` · 219 MB	❌ zip error	⚠️ blank	⚠️ blank	✅ renders _{13.9 s · 4.8 GB}	❌ compose error	⚠️ blank	✅ renders _{1.2 s · 265 MB}
NVIDIA office chair `usdc` · 35 MB	❌ zip error	⚠️ blank	⚠️ blank	✅ renders _{2.1 s · 1.8 GB}	❌ compose error	⚠️ blank	✅ renders _{1.2 s · 231 MB}
imagine.io railing `usdc` · 2 MB	❌ zip error	❌ crash	❌ crash	⚠️ blank	✅ renders _{1.2 s · 292 MB}	⚠️ blank	✅ renders _{1.2 s · 232 MB}

Why they fail

three.js 0.174: USDZLoader only reads zipped .usdz, so every loose .usd / .usda fails with invalid zip data.
three.js 0.186: USDLoader reads binary USDC but only composes layers already in memory. It never fetches referenced files, so multi-file packages render empty. A crawler that prefetches referenced layers fixes Robotiq; NVIDIA's payload chain and Thor stay empty. The railing crashes, likely on its !invert!xformOp:translate:pivot op.
Needle: Pixar's OpenUSD compiled to WASM, and the only library that composes the multi-file packages. It needs COOP/COEP headers (threaded build), grows to ~750 MB of WASM memory before loading anything, and its materials translate inconsistently. The railing binds only MDL materials and renders nothing (likely cause). PolyForm Noncommercial license.
tinyusdz: renders single-file assets; fails with Failed to compose references on NVIDIA's payload chains and renders Robotiq and Thor empty.
cinevva usdjs: renders the self-contained laptop; composes the multi-file packages into 2–6 triangle scenes.
GLB: converted offline with OpenUSD (usd-core) and loaded with three.js GLTFLoader. Renders everything.

Under each rendered cell: time to first frame and peak browser-process memory. Headless Chromium, median of 3 cold loads, files served locally, software rendering (SwiftShader), so treat time and memory as relative, not absolute. Per-cell data size and triangle counts are in the dataset viewer.

Files

Path	Contents
`data/`	The table above as parquet (renders + status per library)
`renders/`	Screenshots and reference thumbnails
`assets/`	The CC-BY-4.0 source packages, unchanged

The railing comes from imagineio/PhysicalAI-SimReady-Assets under CC-BY-NC-4.0, so only its renders are included. All other packages are CC-BY-4.0 from nvidia/simready-assets, LGElectronics/simready-assets, Robotiq-Official/simready-assets and standardbots/simready-thor.

Downloads last month: 278

Number of rows:

Total file size:

320 MB