crutcher / bimm / #58
72%

DEFAULT BRANCH: main
Ran
Jobs 1
Files 51
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 84.779%. Remained the same
#58

push

web-flow 
Crutcher/api review (#64)

* Make `pipeline` module private and remove unused exports in `pipeline/mod.rs`

* Add `try_get` method to improve error handling in `FirehoseRow`

- Introduced `try_get` method to return a `Result` with context for missing columns.
- Simplified `get` implementation by delegating to `try_get`.

* Refactor `FirehoseRow` API: replace `maybe_get` logic with `expect_*` utilities for cleaner column handling and serialized data operations.

* Update dependencies in `Cargo.toml` and `Cargo.lock`

- Bumped versions for `serde_json`, `clap`, `anyhow`, and `proc-macro2`.

* Add argument parsing with `clap` and update `swin_tiny` pipeline configuration

- Added command-line argument parsing to `swin_tiny` using `clap` for batch size, number of workers, and number of epochs.
- Updated patch size and layer configuration for the `swin_tiny` example.
- Adjusted `Cargo.toml` and `Cargo.lock` dependencies accordingly.

* Simplify default value for `num_workers` argument in `swin_tiny` example.

* Switch `lr_scheduler` in `swin_tiny` to Noam and add warmup epochs configuration.

* Add `oversample_ratio` argument and adjust warmup epochs in `swin_tiny` example

- Introduced `oversample_ratio` argument with default value 2.0 to enhance training dataset sampling.
- Reduced default `num_warmup_epochs` value from 2 to 1.
- Updated pipeline configuration to include `oversample_ratio`.

* Refactor `swin_tiny` example: streamline model initialization and adjust warmup steps

- Moved model initialization directly into the learner builder.
- Adjusted warmup step calculation to divide batch size by 2.

* Update defaults in `swin_tiny`: learning rates, warmup epochs, and scheduler

- Increased default `num_warmup_epochs` from 2 to 10.
- Adjusted default `learning_rate` to `1.0e-3` and `min_learning_rate` to `1.0e-5`.
- Revised warmup steps calculation to include model embedding size.

* Increase default `num_warmup_epochs` to 4 and adjust warmup step cal... (continued)

2033 of 2398 relevant lines covered (84.78%)

81.05 hits per line

Jobs
ID Job ID Ran Files Coverage
1 #58.1 51
84.78
Source Files on build #58