Resources

ABSOLUTE TRASH

#17 opened 17 days ago by

joemamabals

[Bug] Official DGX Spark recipe (Marlin + MTP k=3) dies at ~1,076 json_schema requests on GB10 / SM121

#16 opened 26 days ago by

bonellisystems

Update (docs): Included image text-to-test in readme

#15 opened about 2 months ago by

reality-interface

why can not quit with ctrl-c when run with vllm 0.24.0 in jetson agx orin 64G, and MTP generate gibberish

#14 opened 2 months ago by

loveyoupeng

别开MTP，开了变傻逼

#13 opened 3 months ago by

jiecong

Multi-GPU NVFP4 issues on 16 GB cards (RTX 4060 Ti)

➕ 1

#12 opened 3 months ago by

Josephur

TemporalMesh Transformer: 29.4 PPL at 48% compute — beats Mamba, new open-source architecture

#11 opened 3 months ago by

vigneshwar234

DGX Spark is not running qwen3.6 35b a3b nvfp4 on nvidia vllm image

👀 1

#10 opened 4 months ago by

moranilt

[Bug] Three stability failure modes on RTX 5090 (consumer Blackwell, SM120) — illegal instruction, CUDA-graph-replay crash, and silent engine hang

#9 opened 4 months ago by

YoRandom

Bazaar: Monetize your Text-to-Speech Model as a pay-per-query API

#8 opened 4 months ago by

gionuibk

Bazaar: Monetize your Text-to-Speech Model as a pay-per-query API

#7 opened 4 months ago by

gionuibk

Genuinely a fantastic model

👍 1

#5 opened 4 months ago by

swtb

RTX 5090 performence

👍 4

#4 opened 4 months ago by

jasionkajakub

Benchmark Report: Qwen3.6-35B-A3B-NVFP4 on NVIDIA DGX Spark, Jetson Thor, Blackwell 6000 Pro

👍 2

#3 opened 4 months ago by

shakhizat

NVIDIA GeForce RTX 5090

👍 1

#2 opened 4 months ago by

Mosai-Sys

on the DGX spark

👍 5

#1 opened 4 months ago by

shakhizat