torch20 transfer #86
Merged
torch20 transfer #86
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
jgrss
added a commit
that referenced
this pull request
Sep 23, 2024
* fix deleted method * format * format * deprecated scipy * add None check * update Dataset * add registers for collate * minor formatting * remove model tensor reshaping * rename args * network name changes * name changes * add keyword arg * move normalization method * update tests * convert to tensors * update README * format * format * format * format * format * use nn. * format * format * format * format * format * update loss * fix dataset test * move to modules * pin changes * stash * run train session * fix train data storage * add enum * change default transforms to None * change time series start/end get method * update script flow * add tower * add worker check * implement * start on predict data * fix inference steps * replace deprecated * fix predict window * cleanup * set defaults * comment * comment * changes * add foj loss * add foj loss * fix dropout * cleanup params * cleanup transfer * cleanup args * fix deep supervision * fix deep supervision * need to update the spatial gpkg * update dependencies * add 0 check * add multipolygon method * add pool_first arg * add pool_first arg * add pool_first arg * add pool_first arg * add pool_first arg * remove args * script changes * add optional StdConv * add pyogrio * replace gw polygon_to_array * add CLI arg for unknown labels * make read method * update progress bars * fix augmentations * fix augmentations * fix augmentation bdist * add args * add args * transfer updates * move methods * move scale * pass conv args * move methods * move data methods * add module import * modfy CLI defaults * update create methods * replace tqdm with rich * add batch file store * remove old code * support weak supervision * fix tests * test mask-rcnn * stash * change conv args * change conv args * test zoom and support score masking * modify tower unet * change arg name * cleanup unused and add order option * cleanup unused and add order option * change CLI * update CLI args * update test * change module names * stash * stash * add temp exception * reconfigure resconv * remove files * modify transfer model * fix mask scoring * use transfer name * separate spatial splits * add missing method * transfer loss * all_touched=True * add data methods * add edge check * add transfer arg * finetune * re-test params * get utm zone from lat/lon * change time scaler param * pass activation param * experimental final layers * cleanuup refinee * start/end * enums * enums * topology loss * cleanup * test attention * vit * residual in residual * attention * remove CLI option * predict start/end * temp test * libraries * reconfigure augmentations * random rng support * simplify augmenter * add data check for geometry collection * augment probability handling * force new spatial data for transfer model * formatting * test new loss * move modules * move modules * move modules * profiler * module imports * docs: update README (#75) * feat: transfer ltae v2 (#76) * update README * remove import * pass all_touched arg * add lon/lat to data batch * pass persistent workers arg * 👕 formatting * pass lon/lat coordinates * cleanup loss * rename network names * remove ViT * try/except import natten * 👕 formatting * pass all_touched from CLI * add all_touched CLI arg * fix: torch20 tests (#77) * 🔨 modify CI workflow * ⚡ remove install line * ⚡ update augmentation test names * 📝 update torch version in README * 🔨 modify CI workflow * 🔨 setup lib requirements * update augmentation tests * ⚡ update data tests * ⚡ update data tests * ⚡ better einops characters * 👕 formatting * ➕ raise minimum torchmetrics version * 🎨 formatting * fix: remove instances of torch_geometric (#78) * fix: test batch save step (#80) * docs: torch20 docs (#79) * 📝 update README * 📝 add sub-md * 📝 add install line * 📝 remove line * 📝 better comments * 📝 add README links * fix: torch20 cleanup (#81) * 🗑️ remove redundant method * 🎨 formatting * 🔥 py3 super() and remove unused methods * ✅ add tests for transformer * 🎨 formatting * ✅ add tests for tower unet * 🎨 formatting * 🔥 remove FOJ * 🔥 py3 super() and remove unused methods * 🎨 formatting * 🔥 py3 super() and remove unused methods * fix: add fiona dependency (#82) * fix: use cpu device (#83) * 🐛 add supported attention * 🐛 revert attention * 🐛✅ attention tests * 🙈 hide test for now * fix: torch20 train test (#84) * ✏️ change variable name * 🎨 formatting * ✅ test other weights * 🙈 block train CLI test * fix: torch20 network c (#85) * 🔥 remove network c module * 🎨 formatting * torch20 transfer (#86) * 🎨 formatting * 🔥 remove gain from dask storage * 🎨 formatting * 🎨 formatting * ⚡️ change CLI defaults * ✨ new loss enum * ⚡️ relative import * ✅ update tests * 🔒️ upgrade setuptools * fix: torch20 ex (#87) * ⚡️ sum weights * ⚡️ move sum * ✅ update loss tests * 🎨 formatting * ⚡️ use common method * 🎨 formatting * ➖ make kornia optional * ➕ increase dependency versions * ➖ cleanup imports * fix: ➖ remove boundary loss (#88) * ➖ remove topology loss * 📝 add comments for prior change * ⚡️ change default loss * 🎨 * ⚡️ add CLI data-pattern; remove args * 🎨 * 🔥 cleanup unused methods * 📝 new relative imports * ⚡️ change defaults * 🎨 * 🎨 * ⚡️ import loss * 🎨 * 🎨 * 🎨 * 🎨 * ✅ tests * 🎨 * 🎨 * 📝 change var name * 📝 change var name * change to class * ⚡️ predict callback return 3 bands * 🔥 remove unused import * 🎨 * ⚡️ relative loss imports * ⚡️ update enums * 🔥 remove arg in predict method * 🔥 remove temporal transformer * 🎨 * ⚡️ update main model network * ⚡️ ensure grad * 🎨 * 🎨 * 🎨 * 🔥 remove file * ✨ new relative import * 🔥 remove imports * 🎨 * 🔥 remove file * ⚡️ tests * ⚡️ scripts * 🔥 remove keyword arg * ⚡️ add keywrod arg * ⚡️ comment * 🎨 * 🎨 * 🎨 * 🎨 * 🔥 remove code * 🎨 * 🔥 remove imports * 🚑️ comment * 🎨 * 🔥 remove unused modules * 🎨 * 🎨 * 🐛 change default classes * 🔥 remove arg * 📝 * ⚡️ imports * 🐛 better check * ⚡️ change default values * ⚡️ rearrange convolutions * 🎨 * ⚡️ simplify pre * 🔥 remove unused losses * 🔥 remove unused module * 📝 static version --------- Co-authored-by: Ubuntu <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR:
setuptools