|
Gradual host memory growth (no plateau) on Parakeet-CTC offline multi-locale NIM (image 1.5.3, Riva release_version 2.27.0-dev)
|
|
0
|
13
|
September 11, 2026
|
|
AI for Media R22 rel
|
|
0
|
85
|
September 11, 2026
|
|
Parakeet CTC Vietnamese: Is the original ARPA LM available for n-gram LM interpolation / fine-tuning?
|
|
3
|
124
|
September 10, 2026
|
|
Voice AI Developers: What Architecture Delivers the Best Latency, Quality, and Cost Balance?
|
|
0
|
23
|
September 10, 2026
|
|
Intermittent Popping / Crackling Noise with Magpie Multilingual TTS on Jetson AGX Thor
|
|
0
|
14
|
September 8, 2026
|
|
Request for Access: Riva TTS Magpie
|
|
6
|
264
|
September 5, 2026
|
|
NVIDIA RIva Magpie access
|
|
4
|
91
|
September 2, 2026
|
|
BF16 strongly-typed engine build segfaults in Myelin on RTX 5090 (sm_120) — identical network builds fine as FP16; reproduced on TensorRT 11.0 / 11.1
|
|
3
|
73
|
August 31, 2026
|
|
OBSBOT Center crashes with null function-pointer call on graphics thread every time Eye Tracking (AR SDK) is enabled — RTX 5090 Laptop, AR SDK 0.8.7
|
|
0
|
33
|
August 28, 2026
|
|
PyTorch Conditionals that run on GPU / TensorRT
|
|
3
|
89
|
August 28, 2026
|
|
Aliased I/O TensorRT plugin bug
|
|
1
|
39
|
August 28, 2026
|
|
TensorRT 10.14 silently produces wrong detection scores for D-FINE (DETR-style) models on RTX 5090 / sm_120 — default build drops 12 detections to 0,
|
|
2
|
86
|
August 28, 2026
|
|
Model meta/llama-3.1-8b-instruct no longer supported
|
|
0
|
52
|
August 27, 2026
|
|
Held-out group went 72% → 100%. That's the number that actually matters
|
|
0
|
31
|
August 26, 2026
|
|
/generate silently drops forwarded HTTP headers — we're flying blind on multi-tenant LLM traffic (PR #8916 open)
|
|
0
|
29
|
August 24, 2026
|
|
Access to SCRFD-10G Face and Landmarks Detector weights under NVIDIA Open Model License
|
|
0
|
37
|
August 24, 2026
|
|
Porting bit-exact QKV fusion from CPU to TensorRT
|
|
0
|
35
|
August 24, 2026
|
|
[Blackwell] cuDNN fused FP16 SDPA decode kernel (q_len=1) hangs under MPS multi-tenancy
|
|
0
|
34
|
August 21, 2026
|
|
TrafficCamNet Transformer Lite: Fine-tuning from 4 to 9 classes — RT-DETR weights or ResNet-50 backbone weights?
|
|
2
|
47
|
August 19, 2026
|
|
What are good practices for building reliable AI workflows that interact with external APIs?
|
|
2
|
74
|
August 18, 2026
|
|
Yin Yan appled to code
|
|
0
|
19
|
August 18, 2026
|
|
Is 32× A100 40 GB with 1 GPU per node a reasonable distributed-training testbed?
|
|
1
|
31
|
August 18, 2026
|
|
Qwen3.8-27B at 256K on a 24 GB Blackwell target GPU: iMatrix NVFP4 + MTP, 55.4 tok/s
|
|
0
|
381
|
August 18, 2026
|
|
TensorRT inferior scalability on A10, L4 possibly due to locking issues
|
|
8
|
317
|
August 17, 2026
|
|
Where is Qwen 3.8 NVFP4?
|
|
0
|
258
|
August 17, 2026
|
|
DeepSeek-Coder-V2-Lite 16B on RTX 2050 laptop — seeking inference feedback
|
|
3
|
102
|
August 16, 2026
|
|
Work whit agent
|
|
0
|
37
|
August 16, 2026
|
|
Individual developer access to Windows AR SDK Core
|
|
0
|
27
|
August 16, 2026
|
|
No artifact → no claim → exit 1. No hash → no trust. No zip → no history. System records existence, not truth
|
|
0
|
42
|
August 13, 2026
|
|
How to Scale from 2 to 10+ DGX Spark Systems for a ChatGPT-Like AI Platform?
|
|
2
|
170
|
August 10, 2026
|