Melanoma task #310
Melanoma task #310
Conversation
…e. also delete the tarball after done using
Bumps [JamesIves/github-pages-deploy-action](https://github.com/jamesives/github-pages-deploy-action) from 4.6.8 to 4.6.9. - [Release notes](https://github.com/jamesives/github-pages-deploy-action/releases) - [Commits](JamesIves/github-pages-deploy-action@v4.6.8...v4.6.9) --- updated-dependencies: - dependency-name: JamesIves/github-pages-deploy-action dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
R/TaskClassif_melanoma.R
Outdated
| #' | ||
| #' @references | ||
| #' `r format_bib("melanoma2021")` | ||
| #' @examplesIf torch::torch_is_installed() |
There was a problem hiding this comment.
I don't think this is required here, as printing the task does not create any tensors
|
|
||
| compressed_tarball_file_name = "hf_ISIC_2020_small.tar.gz" | ||
| compressed_tarball_path = file.path(path, compressed_tarball_file_name) | ||
| curl::curl_download(paste0(base_url, compressed_tarball_file_name), compressed_tarball_path) |
There was a problem hiding this comment.
because curl is in suggests, we should run mlr3misc::require_namespaces("curl") before so users get a good error message when they don't have it installed.
There was a problem hiding this comment.
But we should just write require_namespaces() without the mlr3misc:: right?
R/TaskClassif_melanoma.R
Outdated
| old = c("image", "patient", "anatom_site_general"), | ||
| new = c("image_name", "patient_id", "anatom_site_general_challenge") | ||
| )[, split := "test"] | ||
| metadata = rbind(training_metadata, test_metadata, fill = TRUE) |
attic/03-process_melanoma.R
Outdated
| @@ -0,0 +1,84 @@ | |||
| library(data.table) | |||
There was a problem hiding this comment.
the attic is more for code that was discarded for now but might be used again in the future (I haven't cleaned up in a while though :D).
If it is about how to create data, I would probably rather put it into data-raw. But as we already document on huggingface how it was created. We don't really need it in this repository I think.
R/TaskClassif_melanoma.R
Outdated
| training_metadata = data.table::fread(here::here(path, training_metadata_file_name)) | ||
|
|
||
| test_metadata_file_name = "ISIC_2020_Test_Metadata.csv" | ||
| test_metadata = data.table::fread(here::here(path, test_metadata_file_name)) |
| old = c("image", "patient", "anatom_site_general"), | ||
| new = c("image_name", "patient_id", "anatom_site_general_challenge") | ||
| )[, split := "test"] | ||
| metadata = rbind(training_metadata, test_metadata) |
There was a problem hiding this comment.
I think fill = TRUE was here because we want to fill the response variable of the test data with NA, will double-check to confirm
There was a problem hiding this comment.
ok makes sense, maybe add a comment
|
And yes, I would delete the individual files on huggingface |
…y_imagenet. Probably need to construct full file paths first
| @@ -83,7 +83,7 @@ load_task_melanoma = function(id = "melanoma") { | |||
| cached_constructor = function(backend) { | |||
| data = cached(constructor_melanoma, "datasets", "melanoma")$data | |||
|
|
|||
| data[, benign_malignant := factor(benign_malignant, levels = c("benign", "malignant"))] | |||
| data[, outcome := factor(outcome, levels = c("benign", "malignant"))] | |||
There was a problem hiding this comment.
the static analyzer will not know that outcome is a valid variable here (this is generally an issue with NSE)
you can use outcome := factor(get("outcome"), ...) as a workaround
https://huggingface.co/datasets/carsonzhang/ISIC_2020_small
Should we delete the individual files on Hugging Face?