Develop: Don't install AVX/AVX2 version of TensorFlow by default #222

Signed-off-by: Michael Mayer <michael@photoprism.app>
This commit is contained in:
Michael Mayer
2025-04-02 11:50:55 +02:00
parent 083c59550a
commit 00e12b8bd8

View File

@@ -123,7 +123,7 @@ services:
# PHOTOPRISM_FFMPEG_BITRATE: "32" # video bitrate limit in Mbit/s (default: 50) # PHOTOPRISM_FFMPEG_BITRATE: "32" # video bitrate limit in Mbit/s (default: 50)
# LIBVA_DRIVER_NAME: "i965" # For Intel architectures Haswell and older which do not support QSV yet but use VAAPI instead # LIBVA_DRIVER_NAME: "i965" # For Intel architectures Haswell and older which do not support QSV yet but use VAAPI instead
## Run/install on first startup (options: update https gpu ffmpeg tensorflow davfs clitools clean): ## Run/install on first startup (options: update https gpu ffmpeg tensorflow davfs clitools clean):
PHOTOPRISM_INIT: "https tensorflow" PHOTOPRISM_INIT: "https"
## Share hardware devices with FFmpeg and TensorFlow (optional): ## Share hardware devices with FFmpeg and TensorFlow (optional):
# devices: # devices:
# - "/dev/dri:/dev/dri" # Intel QSV (Broadwell and later) or VAAPI (Haswell and earlier) # - "/dev/dri:/dev/dri" # Intel QSV (Broadwell and later) or VAAPI (Haswell and earlier)