York Cheung

icon indicating an email [email protected]

icon company UC Irvine icon location California

Results 7 comments of York Cheung

failed to build v0.9.0: ./tensorflow/core/kernels/sparse_matmul_op.h:75:11: error: cannot convert 'const type {aka const float}' to '__vector(4) __builtin_neon_sf'

Oh hey，I'm faced the same ERROR in the same model as yours, when I followed @samjabrahams 's step by step build guide : https://github.com/samjabrahams/tensorflow-on-raspberry-pi/blob/master/GUIDE.md Here's the error information: # bazel...

[Bug]打开应用后只有一个白屏

+1

[BUG/Help] 量化失败：RuntimeError: CUDA Error: no kernel image is available for execution on the device

P100 有同样问题

h2o is slower and not optimized in memory

same here

Image from guidance abnormal

Hello？ Is there anyone can answer me？

Image from guidance abnormal

@andre-nguyen Yes, but later I've got this: ![image](https://cloud.githubusercontent.com/assets/13660547/23603187/eb3c51da-028e-11e7-83c7-8fde585b1e97.png) the data from guidance seems like nothing wrong.. And there's another thing, after I switch the mode from "standard" to "custom" in...

[Bug] RuntimeError: Failed to allocate memory for batch_prefill_tmp_v with size 458752000 and alignment 16 in AlignedAllocator

Same. I use 2 A100, sglang v0.3.1.post3, and cuda graph disabled.