On this video, I discuss what is need to configure the training files. Continued to update the flowchart further define the workflow.
It is them this where we left off. So again this is other attributes that we will have to use you haven’t updated is going on and I’m and I’ll talk about that is going to the application development so goal back here going out here bit so so there will be a decision point here for all this Here is model course of it’s not then it is can go back to the scattering Spartan said there are no general and so here essays no waiting see where reason that will prove it is what’s can happen and is this user Ray here text on a for to schedule training and that pretty much will kickstart will kickstart process so in the background so list of this is not start that so was a kick kickstart it the portal itself in being and then back in the how dashboard changes status as scheduled training and lots model so when the training is scheduled for well when one is actually scheduled for training it’s been a lot that model any new data center reflect that will be really to have nominated the locket so there’s any data out there and that in front and a people roster doing something it’s is not accept known today it’s going on his can of freedom and then what’s can happen after that it’s to start generating on the as a nonfunctional August 2 that that’s what makes year bonanzas just America here nonfunctional user is a working background generates you see files his fees generating the fine and we subsequently generate the label.tv tax and or the company the models.config files as a stick another mystica closer look you guys in those heart fetuses so as a nonfunctional user again this process here will party going the drill down is to have the do it is can a generate the CSV files generated to have record files and execute actually the putting here is flow test/training.record files and labels and configurations models for each and if the model soul after this it’s going to go into some it’s skitter to create a skitter were called skitter is have you processing and and that it’s good it’s been a single to the processes and transition into the next step which will be training so scheduler her cute that were cute that after that’s this this over here is done here so for now we have completely configured the training for the model and we can automate these files here evident in Of a webpage where the function who user can do the object to the next deal that I’ll be discussing is a workflow is this right here execute training now this is what I’m thinking and scope of the schedule is a series of processes or Cron jobs in Windows will be when the scattering is that it can be running 24 seven and is it’s can execute tens of flow training sessions run intensive floor Python trainings and it’s can run until as many staff was required by the configurations models and trainings chess sessions or step and when it’s done it it’s for the explored an inference graph and then move onto the next model in inserts running 24 seven running training models continuous and then no rule we might have a webpage he appeared to see how many of these instances with Tessa flow will will go to discuss his later so we might merge these two steps together as part of the skitter for him will at that sarcasm of meeting on something to anyways here will be for wall reference the I was there gay some of their so in the hit save and will start the next session self-control is school on this next section of the video will talk about the workflow on the configurations so when the do’s and creating new tab here and… Command the next some names in limiting about here so, some things that will see here secured to the flow training someone at the center want to sorry but it’s a lot of there is a general from this sketch of the things that another absolutely there will be function user and I guess because we just can be delaying with configurations think this is going to be fine done this in this adherent are party just enemy mentioning web at the web and print the web portal so goal back to basics low on the start here and you want to do is so reference previous that says ready for her for training so this step will is reference this to labeling and classification so when want to do is in the web portal wanted generates dashboard lives will be to amusement that he you guys. So when this whole process here is the labeling is done is the fact that in minutes it’s marked with Allah labels are down is ready for training the community at that dashboard all the models very ready to be configured so once the one has is another web portal long essay on having a webpage. User can configure training might soul when I meet women whining about this is we look at some of the requirements that we have to he is this this will we had generate this file in order for console you know we’re going to the copy this so this is sound where to look for the second and what could happen is really get this in this is a file that we need so were put into the floor here in use as a reference So this file I could do right now is not in the be generated until the trainings actually executed because the labeling the labels in the be managed during the labeling process so in the database here been a have the stables in the database to the files they beat generator at a later step that his place in the here so what I to switch between Notepad nine here is was a of part a clearing use web portal webpage there is distance here is that some of this configuration in needs needs input from the user for these for the functional user for example somebody like myself that as it is getting the data together form create a model this number of classes is number Again can be generated by the system when Scheduled and everything else is automated the trainings on but is what I what’s have some kind of input will be some of these attributes for example I may choose to do decide the model and some other options will be if I can find it been mounted stats I wanted to train for for example here to this attribute here calling attributes when he was caught attribute has this value right now by default past 200,000 steps but I might be public of party may want to do a model to do a quick train conceive it to work for me I my want to choose it to do hundred thousand training or 500 just quick training and in the the data the train data says to just check I don’t think that you really need to hundred thousand steps a training box the options that some to highlight this is my configurations as use it as an example I might have the option of selecting would like to continue that he made sure to subject is so you guys this capture this nano highlight this copy that is sense known and adhere and should give it in on the group is usually had just to represent and with do this this evening this let me in the change the color to that may be also change the connector to be in one select and norm and the think – just a Caribbean is there was I’ll get one of his these days object the there you back to point I so just a quick to notes on to this year what is the requirements be one for users to be the training will days on training models user can update attributes for so and most likely would party going to details as to how the application with so images and this is just a reference so will be the next these once that is configured there will go back to dashboard here and is the simple but to start start scheduled training so good with the examiner do is a may want to introduce what I want to dashboard but first this go back to this data in an is moving over here are the other thing that we week we had discussed in the previous video is this data is data the guess create difficult test labels and train labels is that the data format this capture that example and for this one will do is able label him and he protection from this this is a strange last as CSV file that I so the next thing they were looking for this another one this comes year is this it’s it’s he is to will phonetic Notepad he was talking about this to do so we have to create this very model so the going here is interesting. You to select this is there is a letter here is that this was his old generating that I missing here so these files the three files here can be generated automatically when the labeling process happened to step in this step number two here all these data can be saved in the database in this format we can track with height of an image he can also track the X and Y minimal indexing Weinman wall-mounted box then of course reference old image locations and create we can create to CSV in other words we don’t necessarily what needs the X amount, not less somebody is doing the XML files in her own maybe you offer some API when it can upload that is a file with the image in XML that would later process is important for the 17 but that’s online and it’s needed will build API but in our case everything is good to be a web portal the labeling for been the web portal built load of images that been a web portal and all the data is good be stored in a database so all we have to do is from the database extract this into a CSV format required for consultant half and again in the old recent picked behind I is because when you generate this TRF record high script all non-do we have to update this if statements here in accordance with the labels and we half for a specific model and us to execute the script in order to create the TRF records train.TF Records and CBS and is looking he has been so and then it’s needed website he represents that it’s good to generate those records second think it’s train.records many something that is making you sooner generate these two out but not these two outputs right here these two files is is what is needed to start the actual training process right so we’re executing these and which again all this can be automated and this talk about that affirmations as we progress through the development of the application on Tuesday and then I see these reference train records and test records are referenced in the configuration of the right here she this test and train records so that will be other attributes and when he a bit of December here all close to this video blog are my personal opinion and they don’t in any way reflect opinions of my employer all materials posts and advice from this site are informational researching and for testing purposes of you can use them at your own responsibilities I not in any way responsible for any damage done by following posts advice tutorials and articles from this blog my machine learning blog man not own some of the continent percent