Isabella Tromba
Isabella Tromba
The overview page contains a description but it's easy to miss and should be on the metrics page as well.
Currently, the python library only takes json as input to make predictions. It would be convenient to also allow pandas data frames as input.
Currrently , the CLI outputs "No such file or directory" but should be more verbose and explain which file could not be found. ``` No such file or directory (os...
Tangram currently does not have support for booleans but when passing a json object with boolean values, we really should be coercing them to their `enum` values instead of erroring....
Currently, if you omit early stopping options, they are assumed to be None. Preferably, we automatically perform early_stopping but this requires disambiguating between None and key missing in the config....
Currently, in order to compare models in a repo, we need to click around to open up both models side x side. It would be better if we were able...
Improve the UI so we can see all of the columns on the training stats table in the app
``` error: panicked at 'called `Option::unwrap()` on a `None` value', crates/core/stats.rs:365:48 0: backtrace::capture::Backtrace::new 1: tangram::train::train::{{closure}} 2: std::panicking::rust_panic_with_hook 3: std::panicking::begin_panic_handler::{{closure}} 4: std::sys_common::backtrace::__rust_end_short_backtrace 5: _rust_begin_unwind 6: core::panicking::panic_fmt 7: core::panicking::panic 8: ::fold 9:...