Hi @<1671689437261598720:profile|FranticWhale40>
Are you positive the Triton container finished syncing ?
Could you provide the docker log (both the serving and the triton)?
What is the clearml-serving version you are using ?
Could you add a print in the "preprocess" function, just to validate you are getting to the correct model version ?
and here is the serving inference logs:
ffmpeg version 4.3.6-0+deb11u1 Copyright (c) 2000-2023 the FFmpeg developers
built with gcc 10 (Debian 10.2.1-6)
configuration: --prefix=/usr --extra-version=0+deb11u1 --toolchain=hardened --libdir=/usr/lib/x86_64-linux-gnu --incdir=/usr/include/x86_64-linux-gnu --arch=amd64 --enable-gpl --disable-stripping --enable-avresample --disable-filter=resample --enable-gnutls --enable-ladspa --enable-libaom --enable-libass --enable-libbluray --enable-libbs2b --enable-libcaca --enable-libcdio --enable-libcodec2 --enable-libdav1d --enable-libflite --enable-libfontconfig --enable-libfreetype --enable-libfribidi --enable-libgme --enable-libgsm --enable-libjack --enable-libmp3lame --enable-libmysofa --enable-libopenjpeg --enable-libopenmpt --enable-libopus --enable-libpulse --enable-librabbitmq --enable-librsvg --enable-librubberband --enable-libshine --enable-libsnappy --enable-libsoxr --enable-libspeex --enable-libsrt --enable-libssh --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvorbis --enable-libvpx --enable-libwavpack --enable-libwebp --enable-libx265 --enable-libxml2 --enable-libxvid --enable-libzmq --enable-libzvbi --enable-lv2 --enable-omx --enable-openal --enable-opencl --enable-opengl --enable-sdl2 --enable-pocketsphinx --enable-libmfx --enable-libdc1394 --enable-libdrm --enable-libiec61883 --enable-chromaprint --enable-frei0r --enable-libx264 --enable-shared
libavutil 56. 51.100 / 56. 51.100
libavcodec 58. 91.100 / 58. 91.100
libavformat 58. 45.100 / 58. 45.100
libavdevice 58. 10.100 / 58. 10.100
libavfilter 7. 85.100 / 7. 85.100
libavresample 4. 0. 0 / 4. 0. 0
libswscale 5. 7.100 / 5. 7.100
libswresample 3. 7.100 / 3. 7.100
libpostproc 55. 7.100 / 55. 7.100
Input #0, mov,mp4,m4a,3gp,3g2,mj2, from '2023-12-20_23-29-08-front.mp4':
Metadata:
major_brand : mp42
minor_version : 0
compatible_brands: mp42mp41isomiso2
creation_time : 2023-12-20T23:28:14.000000Z
Duration: 00:01:00.08, start: 0.000000, bitrate: 3517 kb/s
Stream #0:0(und): Video: h264 (High) (avc1 / 0x31637661), yuv420p, 1280x960, 3514 kb/s, SAR 1:1 DAR 4:3, 24.04 fps, 24 tbr, 10k tbn, 20k tbc (default)
Metadata:
creation_time : 2023-12-20T23:28:14.000000Z
handler_name : VideoHandler
Stream mapping:
Stream #0:0 -> #0:0 (h264 (native) -> mjpeg (native))
[swscaler @ 0x5592da50a780] deprecated pixel format used, make sure you did set range correctly
Output #0, image2, to '2023-12-20_23-29-08-front/%04d.jpg':
Metadata:
major_brand : mp42
minor_version : 0
compatible_brands: mp42mp41isomiso2
encoder : Lavf58.45.100
Stream #0:0(und): Video: mjpeg, yuvj420p(pc), 1280x960 [SAR 1:1 DAR 4:3], q=2-31, 200 kb/s, 12 fps, 12 tbn, 12 tbc (default)
Metadata:
creation_time : 2023-12-20T23:28:14.000000Z
handler_name : VideoHandler
encoder : Lavc58.91.100 mjpeg
Side data:
cpb: bitrate max/min/avg: 0/0/200000 buffer size: 0 vbv_delay: N/A
frame= 721 fps=177 q=24.8 Lsize=N/A time=00:01:00.08 bitrate=N/A speed=14.8x
video:26746kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: unknown
ffmpeg version 4.3.6-0+deb11u1 Copyright (c) 2000-2023 the FFmpeg developers
built with gcc 10 (Debian 10.2.1-6)
configuration: --prefix=/usr --extra-version=0+deb11u1 --toolchain=hardened --libdir=/usr/lib/x86_64-linux-gnu --incdir=/usr/include/x86_64-linux-gnu --arch=amd64 --enable-gpl --disable-stripping --enable-avresample --disable-filter=resample --enable-gnutls --enable-ladspa --enable-libaom --enable-libass --enable-libbluray --enable-libbs2b --enable-libcaca --enable-libcdio --enable-libcodec2 --enable-libdav1d --enable-libflite --enable-libfontconfig --enable-libfreetype --enable-libfribidi --enable-libgme --enable-libgsm --enable-libjack --enable-libmp3lame --enable-libmysofa --enable-libopenjpeg --enable-libopenmpt --enable-libopus --enable-libpulse --enable-librabbitmq --enable-librsvg --enable-librubberband --enable-libshine --enable-libsnappy --enable-libsoxr --enable-libspeex --enable-libsrt --enable-libssh --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvorbis --enable-libvpx --enable-libwavpack --enable-libwebp --enable-libx265 --enable-libxml2 --enable-libxvid --enable-libzmq --enable-libzvbi --enable-lv2 --enable-omx --enable-openal --enable-opencl --enable-opengl --enable-sdl2 --enable-pocketsphinx --enable-libmfx --enable-libdc1394 --enable-libdrm --enable-libiec61883 --enable-chromaprint --enable-frei0r --enable-libx264 --enable-shared
libavutil 56. 51.100 / 56. 51.100
libavcodec 58. 91.100 / 58. 91.100
libavformat 58. 45.100 / 58. 45.100
libavdevice 58. 10.100 / 58. 10.100
libavfilter 7. 85.100 / 7. 85.100
libavresample 4. 0. 0 / 4. 0. 0
libswscale 5. 7.100 / 5. 7.100
libswresample 3. 7.100 / 3. 7.100
libpostproc 55. 7.100 / 55. 7.100
Input #0, mov,mp4,m4a,3gp,3g2,mj2, from '2023-12-20_23-29-08-front.mp4':
Metadata:
major_brand : mp42
minor_version : 0
compatible_brands: mp42mp41isomiso2
creation_time : 2023-12-20T23:28:14.000000Z
Duration: 00:01:00.08, start: 0.000000, bitrate: 3517 kb/s
Stream #0:0(und): Video: h264 (High) (avc1 / 0x31637661), yuv420p, 1280x960, 3514 kb/s, SAR 1:1 DAR 4:3, 24.04 fps, 24 tbr, 10k tbn, 20k tbc (default)
Metadata:
creation_time : 2023-12-20T23:28:14.000000Z
handler_name : VideoHandler
Stream mapping:
Stream #0:0 -> #0:0 (h264 (native) -> mjpeg (native))
[swscaler @ 0x55d076d5d740] deprecated pixel format used, make sure you did set range correctly
Output #0, image2, to '2023-12-20_23-29-08-front/%04d.jpg':
Metadata:
major_brand : mp42
minor_version : 0
compatible_brands: mp42mp41isomiso2
encoder : Lavf58.45.100
Stream #0:0(und): Video: mjpeg, yuvj420p(pc), 1280x960 [SAR 1:1 DAR 4:3], q=2-31, 200 kb/s, 12 fps, 12 tbn, 12 tbc (default)
Metadata:
creation_time : 2023-12-20T23:28:14.000000Z
handler_name : VideoHandler
encoder : Lavc58.91.100 mjpeg
Side data:
cpb: bitrate max/min/avg: 0/0/200000 buffer size: 0 vbv_delay: N/A
frame= 721 fps=163 q=24.8 Lsize=N/A time=00:01:00.08 bitrate=N/A speed=13.6x
video:26746kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: unknown
/usr/local/lib/python3.9/site-packages/pyparsing/core.py:854: RuntimeWarning: coroutine 'Preprocess._process' was never awaited
loc, tokens = self.parseImpl(instring, pre_loc, doActions)
RuntimeWarning: Enable tracemalloc to get the object allocation traceback
Yes I'm sure that the Triton container finished syncing.
here is the Triton logs:
I0223 15:58:32.515979 71 model_repository_manager.cc:1352] successfully loaded 'yolo_2' version 1
I0223 15:58:32.842511 71 model_repository_manager.cc:1352] successfully loaded 'yolo_1' version 1
I0223 15:58:32.842579 71 server.cc:559]
+------------------+------+
| Repository Agent | Path |
+------------------+------+
+------------------+------+
I0223 15:58:32.842606 71 server.cc:586]
+-------------+-----------------------------------------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------+
| Backend | Path | Config |
+-------------+-----------------------------------------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------+
| onnxruntime | /opt/tritonserver/backends/onnxruntime/libtriton_onnxruntime.so | {"cmdline":{"auto-complete-config":"true","min-compute-capability":"6.000000","backend-directory":"/opt/tritonserver/backends","default-max-batch-size":"4"}} |
+-------------+-----------------------------------------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------+
I0223 15:58:32.842629 71 server.cc:629]
+---------+---------+--------+
| Model | Version | Status |
+---------+---------+--------+
| yolo_1 | 1 | READY |
| yolo_2 | 1 | READY |
+---------+---------+--------+
I0223 15:58:32.869848 71 metrics.cc:650] Collecting metrics for GPU 0: NVIDIA GeForce RTX 3080
I0223 15:58:32.871196 71 tritonserver.cc:2176]
+----------------------------------+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| Option | Value |
+----------------------------------+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| server_id | triton |
| server_version | 2.24.0 |
| server_extensions | classification sequence model_repository model_repository(unload_dependents) schedule_policy model_configuration system_shared_memory cuda_shared_memory binary_tensor_data statistics trace |
| model_repository_path[0] | /models |
| model_control_mode | MODE_POLL |
| strict_model_config | 0 |
| rate_limit | OFF |
| pinned_memory_pool_byte_size | 268435456 |
| cuda_memory_pool_byte_size{0} | 67108864 |
| response_cache_byte_size | 0 |
| min_supported_compute_capability | 6.0 |
| strict_readiness | 1 |
| exit_timeout | 30 |
Also I can’t call the “preprocess” function since there is no valid endpoint to be hitting
I'm using the latest version of clearml-serving
Name: clearml-serving
Version: 1.3.0
Also I can’t call the “preprocess” function since there is no valid endpoint to be hitting
Wait now I'm confused, when you are calling " None " you are actually calling the preprocess function running on the inference container, and this one in turn (automatically) calls the Triton container.
Are you calling the Triton manually?
Could you share your preprcoess.py , and the command line you have used to register the two model versions ?
(based on your logs everything seems to be loaded correctly, hence my question)
@<1671689437261598720:profile|FranticWhale40> I might have found something, let me see if I can reproduce it
Thanks @<1671689437261598720:profile|FranticWhale40> !
I was able to locate the issue, fix should be released later today (or worst case tomorrow)
@<1671689437261598720:profile|FranticWhale40> could you test the fix? just pull & run
allegroai/clearml-serving-triton:1.3.1
allegroai/clearml-serving-inference:1.3.1
@<1523701205467926528:profile|AgitatedDove14> Also could you please share the commit to fix the issue? It'll help to address it on our end.
Thanks!
@<1671689437261598720:profile|FranticWhale40> this one: None
Thanks @<1523701205467926528:profile|AgitatedDove14> , this seems to solve the issue. I guess the main issue is that the delimiter is a _ instead of / . This did work, however, as you can see from the model endpoint deployment snippet, we also provide a custom aux-config file. We also had to make sure to update the name inside config.pbtxt so that Triton is happy:
From
name: "mmdet"
TO:
name: "mmdet_VERSION" -> "mmdet_1"
we also provide a custom
aux-config
file. We also had to make sure to update the name inside
config.pbtxt
so that Triton is happy:
Good point, what would be the logic of the auto "config.pbtxt" patching we should employ ?
If you’re wondering about the case where no optional config.pbtxt is provided, I guess the logic would be pretty much the same as above:
model_name = f"{model_name}_{version}"
But then after looking at create_config_pbtxt() , it seems like this is not being constructed at all, making me realize that this may have been optional - confirming name is an optional propery . It may just be simpler to drop this value all together.
Or to be explicit, something like this:
config_dict.put("name", f"{endpoint.name}_{endpoint.version}")
making me realize that this may have been optional
I think it is optional, and this is why it was not entered in the first place.
Could you double check and just remove it from your manual pbtxt ?
Sure,
clearml-serving --id $SERVING_ID model add \
--name "yolo" --version 2 --project $PROJECT_NAME --engine triton --endpoint "yolov8" \
--preprocess "./yolo/preprocess.py" \
--input-size 3 -1 -1 --input-name "images" --input-type float32 \
--output-size -1 -1 --output-name "output0" --output-type float32 \
--aux-config ./yolo/config.pbtxt