doing something
712 subscribers
910 photos
36 videos
36 files
1.94K links
@smlkw doing something, just my notes to keep updates
Download Telegram
doing something
CUDA execution
Idk why, FP16 execution is slower than FP32 shown on this picture
Forwarded from A. S.
😁7
Google Colab has been updated
Made a workable program in Rust to do automatic speech recognition of wav2vec2 models with IREE runtime

#rust #ai #audio #speech
IREE runtime with wav2vec2 models (FP32) on CUDA and CPU devices

#iree #rust #audio
Testing Gemma 3 with 12 billion parameters in an image understanding task on H200 card

vLLM used 100 GB of memory

#cv #llms #vllm #cloud
who knows how to delete this unused software?
Forwarded from Tech Crimes (Tailen)