nf-core/deepmodeloptim
Stochastic Testing and Input Manipulation for Unbiased Learning Systems
Define where the pipeline should find input data.
Input data CSV file
stringthe input file containing all input data
Data config YAML
stringthe json config file that specifies all the parameters relative to the data manipulation.
Model file in Python
stringthe model file in python, the model that will be tested by this pipeline.
Model config YAML
stringthe config file with all the hyperparameter directives (choiches) and all ray tune specs.
The output directory where the results will be saved. You have to use absolute paths to storage on Cloud infrastructure.
stringThe directory will contain a subdirectory with a name unique to each stimulus pipeline run.
flag to save transformed csv data.If set to true, the transformed data will be saved in the output directory.
booleanflag to save transformed csv data.
Email address for completion summary.
string^([a-zA-Z0-9_\-\.]+)@([a-zA-Z0-9_\-\.]+)\.([a-zA-Z]{2,5})$Set this parameter to your e-mail address to get a summary e-mail with details of the run sent to you when the workflow exits. If set in your user config file (~/.nextflow/config) then you don't need to specify this on the command line for every run.
Files that can be provided as optional input or omitted.
File to be used to initialize the model in tuning
stringThe initial weights of the model. These files can be used to start the training instead of random initialization. One can provide several files, each of them will be used for a different run.
Options for preprocessing the data
Preprocessing config YAML
stringthe config file that specifies all the parameters relative to the data manipulation.
Name of iGenomes reference.
stringIf using a reference genome configured in the pipeline using iGenomes (not recommended), use this parameter to give the ID for the reference. This is then used to build the full paths for all required reference genome files e.g. --genome GRCh38.
See the nf-core website docs for more details.
The base path to the igenomes reference files
strings3://ngi-igenomes/igenomes/Size of the peak extracted from the BED file
integer300the size of the peak extracted from the BED file.
Options for tuning the model
Range of trials for the tuning. Defined as min,max,step.
stringthe range of trials for the tuning. Defined as min,max,step. e.g. 5,10,5. If not provided the value will be extracted from the provided stimulus config.
Number of replicates for the tuning.
integer1the number of replicates for the tuning, that is how many times the tuning with the same model is run on the same data.
Options for evaluating the model
Prediction data CSV file
stringthe input file containing all input data for the prediction.
Specify maximun processes resources
set maximum GPU limit
integer1requesting the gpus for the tuning steps.
Container for development stage
stringContainer for development stage.
What to do and how to handle errors
Tells the pipeline how to behave on error
stringfinishrefer to nextflow errorStrategy documentation for more details.
number of time to retry if err_strat is set to retry
integerthis also acts as a multiplier for recources request. If it failed for lack of resources it automaticly asks more the second time. take a look at test.conf for more details.
options to skip or change bhaviour of pipeline
checks if all input are comatible and the model can be tuned.
booleantrueflag to tell whether to check or not if the model can be tuned and trained. It does one call of the batch function, (predicting), of the model importing and using everything needed for that. Default run such a check.
optional flag to do a more/less extensive check during check_model.
stringThis will override user given num_sample value for the tune run. This will give the user control on how extensive it wants the check to be. by default is going to be set to 3.
run the shuffle sanity check
booleantrueflag to tell wether to shuffle or not the data and run a train on it. Sanity check always run on default.
developer flag
booleanflag used to switch to debug mode for the pipeline. more verbose outputs.
Parameters used to describe centralised config profiles. These should not be edited.
Git commit id for Institutional configs.
stringmasterBase directory for Institutional configs.
stringhttps://raw.githubusercontent.com/nf-core/configs/masterIf you're running offline, Nextflow will not be able to fetch the institutional config files from the internet. If you don't need them, then this is not a problem. If you do need them, you should download the files from the repo and tell Nextflow where to find them with this parameter.
Institutional config name.
stringInstitutional config description.
stringInstitutional config contact information.
stringInstitutional config URL link.
stringLess common options for the pipeline, typically set in a config file.
Display version and exit.
booleanMethod used to save pipeline results to output directory.
stringThe Nextflow publishDir option specifies which intermediate files should be saved to the output directory. This option tells the pipeline what method should be used to move these files. See Nextflow docs for details.
Email address for completion summary, only when pipeline fails.
string^([a-zA-Z0-9_\-\.]+)@([a-zA-Z0-9_\-\.]+)\.([a-zA-Z]{2,5})$An email address to send a summary email to when the pipeline is completed - ONLY sent if the pipeline does not exit successfully.
Send plain-text email instead of HTML.
booleanDo not use coloured log outputs.
booleanIncoming hook URL for messaging service
stringIncoming hook URL for messaging service. Currently, MS Teams and Slack are supported.
Boolean whether to validate parameters against the schema at runtime
booleantrueBase URL or local path to location of pipeline test dataset files
stringhttps://raw.githubusercontent.com/nf-core/test-datasets/Suffix to add to the trace report filename. Default is the date and time in the format yyyy-MM-dd_HH-mm-ss.
string