128d9729c3
Update davit.py
3 years ago
143a239e93
Update davit.py
3 years ago
5f0a9e9a9f
Update davit.py
3 years ago
6abb3ba634
Update davit.py
3 years ago
e8fd52e39f
Update davit.py
3 years ago
c902731699
Update davit.py
3 years ago
2a55d9516e
Update davit.py
3 years ago
3a422f735a
Update davit.py
3 years ago
e5847a93c4
Update davit.py
3 years ago
cf1cc8b1d1
Update davit.py
3 years ago
0ed0e9ac35
Update davit.py
3 years ago
6692983832
Update davit.py
3 years ago
62eb0c4346
Update davit.py
3 years ago
c8559878fb
Update davit.py
3 years ago
6e5be758dc
Update davit.py
3 years ago
385a3bdb5e
Update davit.py
3 years ago
6521662d58
Update davit.py
3 years ago
284de330cd
Update davit.py
3 years ago
d7930c70bd
update
3 years ago
92f43964de
Create davit.py
3 years ago
3785c234d7
Remove clip vit models that won't be ft and comment two that aren't uploaded yet
3 years ago
755570e2d6
Rename _pretrained.py -> pretrained.py, not feasible to change the other files to same scheme without breaking uses
3 years ago
72cfa57761
Add ported Tensorflow MaxVit weights. Add a few more CLIP ViT fine-tunes. Tweak some model tag names. Improve model tag name sorting. Update HF hub push config layout.
3 years ago
4d5c395160
MaxVit, ViT, ConvNeXt, and EfficientNet-v2 updates
...
* Add support for TF weights and modelling specifics to MaxVit (testing ported weights)
* More fine-tuned CLIP ViT configs
* ConvNeXt and MaxVit updated to new pretrained cfgs use
* EfficientNetV2, MaxVit and ConvNeXt high res models use squash crop/resize
3 years ago
9da7e3a799
Add crop_mode for pretraind config / image transforms. Add support for dynamo compilation to benchmark/train/validate
3 years ago
b2b6285af7
Add two more FT clip weights
3 years ago
5895056dc4
Add openai b32 ft
3 years ago
9dea5143d5
Adding more clip ft variants
3 years ago
444dcba4ad
CLIP B16 12k weights added
3 years ago
dff4717cbf
Add clip b16 384x384 finetunes
3 years ago
883fa2eeaa
Add fine-tuned B/16 224x224 in1k clip models
3 years ago
9a3d2ac2d5
Add latest CLIP ViT fine-tune pretrained configs / model entrypt updates
3 years ago
42bbbddee9
Add missing model config
3 years ago
def68befa7
Updating vit model defs for mult-weight support trial (vit first). Prepping for CLIP (laion2b and openai) fine-tuned weights.
3 years ago
0dadb4a6e9
Initial multi-weight support, handled so old pretraing config handling co-exists with new tags.
3 years ago
9b114754db
refactor push_to_hub helper
3 years ago
ae0a0db7de
Create repo before cloning with Repository.clone_from
3 years ago
803254bb40
Fix spacing misalignment for fast norm path in LayerNorm modules
3 years ago
6635bc3f7d
Merge pull request #1479 from rwightman/script_cleanup
...
Train / val script enhancements, non-GPU (ie CPU) device support, HF datasets support, TFDS/WDS dataloading improvements
3 years ago
0e6023f032
Merge pull request #1381 from ChristophReich1996/master
...
Fix typo in PositionalEncodingFourier
3 years ago
66f4af7090
Merge remote-tracking branch 'origin/master' into script_cleanup
3 years ago
9914f744dc
Add more maxxvit weights includ ConvNeXt conv block based experiments.
3 years ago
8fda68aff6
Fix repo id bug
...
This to fix this issue #1482
3 years ago
1199c5a1a4
clip_laion2b models need 1e-5 eps for LayerNorm
3 years ago
e858912e0c
Add brute-force checkpoint remapping option
3 years ago
b293dfa595
Add CL SE module
3 years ago
a383ef99f5
Make huggingface_hub necessary if it's the only source for a pretrained weight
3 years ago
e069249a2d
Add hf hub entries for laion2b clip models, add huggingface_hub dependency, update some setup/reqs, torch >= 1.7
3 years ago
9d65557be3
Fix errant import
3 years ago
9709dbaaa9
Adding support for fine-tune CLIP LAION-2B image tower weights for B/32, L/14, H/14 and g/14. Still WIP
3 years ago
Fredo Guan