| 23:49 | seku_ | left the channel |
| 23:49 | seku_ | joined the channel |
| 23:57 | johnathan | left the channel |
| 01:14 | comradekingu1 | joined the channel |
| 01:16 | comradekingu | left the channel |
| 01:55 | Bertl | off to bed now ... have a good one everyone!
|
| 01:55 | Bertl | changed nick to: Bertl_zZ
|
| 03:39 | slikdigit | left the channel |
| 03:59 | Spirit532 | left the channel |
| 05:03 | Spirit532 | joined the channel |
| 06:46 | BAndiT1983|away | changed nick to: BAndiT1983
|
| 08:58 | Bertl_zZ | changed nick to: Bertl
|
| 08:58 | Bertl | morning folks!
|
| 09:15 | se6astian|away | changed nick to: se6astian
|
| 09:42 | siddhantsahu | Good Morning!
|
| 09:43 | se6astian | hi siddhantsahu
|
| 09:43 | siddhantsahu | se6astian: how can task T987 : Try running Tensorflow DeepLab-v3+ inside AXIOM Beta help?
|
| 09:44 | siddhantsahu | I mean what is the need of semantic segmentation in AXIOM Beta?
|
| 09:45 | BAndiT1983 | to recognize object and beautify them, have you never used instagram filters? ;)
|
| 09:45 | BAndiT1983 | *objects
|
| 09:49 | RexOrCine | changed nick to: RexOrCine|away
|
| 09:51 | se6astian | siddhantsahu: there is no need currently, its purley a research and experiment centered task
|
| 09:54 | siddhantsahu | se6astian: ok. Let me give it try. Which processor does AXIOM has and how much main memory is available?
|
| 09:55 | BAndiT1983 | zynq
|
| 09:55 | se6astian | 2 arm cores @ 800mhz
|
| 09:55 | BAndiT1983 | not much memory around 1gb, if i remember correctly
|
| 09:55 | se6astian | microzed has 1gb ram in total, shared with fpga
|
| 09:55 | BAndiT1983 | plus the rest of linux system which requires its share
|
| 09:58 | siddhantsahu | These deep learning models are pretty big not sure if the DeepLab-v3+ model fit in the available memory
|
| 09:59 | BAndiT1983 | tensorflow lite maybe
|
| 09:59 | BAndiT1983 | but this is task really not of importance at the moment, as firmware is priority 1
|
| 09:59 | siddhantsahu | BAndiT1983: Yes
|
| 10:00 | siddhantsahu | I know a bit of deep learning, that is why I was asking about this task
|
| 10:02 | BAndiT1983 | can't think of any usage cases at the moment, but i'm a developer and not camera user
|
| 10:15 | Bertl | siddhantsahu: what amount of memory are we talking here?
|
| 10:16 | siddhantsahu | Bertl: Let me check the model details
|
| 10:21 | vup2 | i think without any acceleration from the fpga side semantic segmentation would be nearly impossible
|
| 10:21 | vup2 | but a fun application would be for example to automatically blur any faces to increase anonymity
|
| 10:39 | siddhantsahu | The MobileNetV2 (focused towards mobile devices)model can be used as a feature extractor with DeepLab-v3 which has less precision and number of parameters suited for mobile devices.
|
| 10:39 | Bertl | vup2: yeah, movie makers will love that! :)
|
| 10:42 | siddhantsahu | Refer Table 3 for the MobileNetV2 paper for model size and memory requirements. http://openaccess.thecvf.com/content_cvpr_2018/papers/Sandler_MobileNetV2_Inverted_Residuals_CVPR_2018_paper.pdf
|
| 11:11 | vup2 | It would certainly be a unique feature
|
| 11:40 | BAndiT1983 | se6astian, the menu of axiom remote is now moving when using the virtual knob, but i need some info on how it works for the real one
|
| 11:41 | BAndiT1983 | according to info from Bertl, the knob is using values from 0 to 255 and does also overflow or underflow for this ones
|
| 11:42 | BAndiT1983 | but if i start the visualizer and turn from default state 0 to 1, the menu entries are iterated fast till the end
|
| 11:42 | BAndiT1983 | if i turn back, then back to the top, probably i'm missing something which handles detents of the knob
|
| 11:42 | se6astian | let me dig out the corresponding code lines
|
| 11:42 | BAndiT1983 | ah, found it in main.c, you are subtracting new and last state
|
| 11:43 | BAndiT1983 | int8_t diff = data[0] - qe[0];
|
| 11:43 | BAndiT1983 | knob_event_handler(E1_ROT, diff);
|
| 11:43 | se6astian | yeah thats it
|
| 11:44 | BAndiT1983 | ok, will add corresponding code and upload working version
|
| 11:54 | BAndiT1983 | it works, will commit now
|
| 11:57 | se6astian | great
|
| 11:57 | BAndiT1983 | done, bassic version for now, soem code will be moved to other areas, like init of white balance options, just added it to visualizer for quicker results
|
| 12:17 | slikdigit | joined the channel |
| 12:39 | comradekingu1 | left the channel |
| 13:00 | slikdigit | left the channel |
| 13:00 | slikdigit | joined the channel |
| 13:12 | Bertl | off for now ... bbl
|
| 13:13 | Bertl | changed nick to: Bertl_oO
|
| 14:06 | Spirit532 | left the channel |
| 14:06 | Spirit532 | joined the channel |
| 14:54 | comradekingu | joined the channel |
| 16:52 | anjan | joined the channel |
| 17:42 | BAndiT1983 | changed nick to: BAndiT1983|away
|
| 17:51 | anjan | left the channel |
| 18:34 | slikdigit | left the channel |
| 19:24 | se6astian | off to bed
|
| 19:24 | se6astian | good night
|
| 19:24 | se6astian | changed nick to: se6astian|away
|
| 20:11 | BAndiT1983|away | changed nick to: BAndiT1983
|
| 20:31 | aombk2 | joined the channel |
| 20:32 | aombk | left the channel |