Commit Graph

838 Commits (a5b01ec04e7ba78d0b5ab5c3f2f43a356562a130)

Author SHA1 Message Date
Benjamin Bossan a5b01ec04e Add type annotations to _registry.py
2 years ago
Ross Wightman 4d9c3ae2fb Add laion2b 320x320 ConvNeXt-Large CLIP weights
2 years ago
Ross Wightman d0b45c9b4d Make safetensor import option for now. Improve avg/clean checkpoints ext handling a bit (more consistent).
2 years ago
Ross Wightman 7d9e321b76 Improve tracing of window attn models with simpler reshape logic
2 years ago
Ross Wightman 2e38d53dca Remove dead line
2 years ago
Ross Wightman f77c04ff36 Torchscript fixes/hacks for rms_norm, refactor ParallelScalingBlock with manual combination of input projections, closer paper match
2 years ago
Ross Wightman 122621daef Add Final annotation to attn_fas to avoid symbol lookup of new scaled_dot_product_attn fn on old PyTorch in jit
2 years ago
Ross Wightman 621e1b2182 Add ideas from 'Scaling ViT to 22-B Params', testing PyTorch 2.0 fused F.scaled_dot_product_attention impl in vit, vit_relpos, maxxvit / coatnet.
2 years ago
testbot a09d403c24 changed warning to info
2 years ago
testbot 8470e29541 Add support to load safetensors weights
2 years ago
Ross Wightman 624266148d Remove unused imports from _hub helpers
2 years ago
Ross Wightman 2cfff0581b Add grad_checkpointing support to features_only, test in EfficientDet.
2 years ago
Ross Wightman 9c14654a0d Improve support for custom dataset label name/description through HF hub export, via pretrained_cfg
2 years ago
Ross Wightman 0d33127df2 Add 384x384 convnext_large_mlp laion2b fine-tune on in1k
2 years ago
Ross Wightman 7a0bd095cb Update model prune loader to use pkgutil
2 years ago
Ross Wightman 13acac8c5e Update head metadata for effformerv2
2 years ago
Ross Wightman 8682528096 Add first conv metadata for efficientformer_v2
2 years ago
Ross Wightman 72fba669a8 is_scripting() guard on checkpoint_seq
2 years ago
Ross Wightman 95ec255f7f Finish timm mode api for efficientformer_v2, add grad checkpointing support to both efficientformers
2 years ago
Ross Wightman 9d03c6f526 Merge remote-tracking branch 'origin/main' into levit_efficientformer_redux
2 years ago
Ross Wightman 086bd55a94 Add EfficientFormer-V2, refactor EfficientFormer and Levit for more uniformity across the 3 related arch. Add features_out support to levit conv models and efficientformer_v2. All weights on hub.
2 years ago
Ross Wightman 2cb2699dc8 Apply fix from #1649 to main
2 years ago
Ross Wightman b3042081b4 Add laion -> in1k fine-tuned base and large_mlp weights for convnext
2 years ago
Ross Wightman 316bdf8955 Add mlp head support for convnext_large, add laion2b CLIP weights, prep fine-tuned weight tags
2 years ago
Ross Wightman 6f28b562c6 Factor NormMlpClassifierHead from MaxxViT and use across MaxxViT / ConvNeXt / DaViT, refactor some type hints & comments
2 years ago
Ross Wightman 9a53c3f727 Finalize DaViT, some formatting and modelling simplifications (separate PatchEmbed to Stem + Downsample, weights on HF hub.
2 years ago
Fredo Guan fb717056da Merge remote-tracking branch 'upstream/main'
2 years ago
Ross Wightman 64667bfa0e Add 'gigantic' vit clip variant for feature extraction and future fine-tuning
2 years ago
Ross Wightman 36989cfae4 Factor out readme generation in hub helper, add more readme fields
2 years ago
Ross Wightman 32f252381d Change order of checkpoitn filtering fn application in builder, try dict, model variant first
2 years ago
Ross Wightman bed350f5e5 Push all MaxxViT weights to HF hub, cleanup impl, add feature map extraction support and prompote to 'std' architecture. Fix norm head for proper embedding / feat map output. Add new in12k + ft 1k weights.
2 years ago
Ross Wightman ca38e1e73f Update ClassifierHead module, add reset() method, update in_chs -> in_features for consistency
2 years ago
Ross Wightman 8ab573cd26 Add convnext_tiny and convnext_small 384x384 fine-tunes of in12k weights, fix pool size for laion CLIP convnext weights
2 years ago
Fredo Guan 81ca323751
Davit update formatting and fix grad checkpointing (#7)
2 years ago
Ross Wightman e9aac412de Correct mean/std for CLIP convnexts
2 years ago
Ross Wightman 42bd8f7bcb Add convnext_base CLIP image tower weights for fine-tuning / features
2 years ago
Ross Wightman a2c14c2064 Add tiny/small in12k pretrained and fine-tuned ConvNeXt models
2 years ago
Ross Wightman 2e83bba142 Revert head norm changes to ConvNeXt as it broke some downstream use, alternate workaround for fcmae weights
2 years ago
Ross Wightman 1825b5e314 maxxvit type
2 years ago
Ross Wightman 5078b28f8a More kwarg handling tweaks, maxvit_base_rw def added
2 years ago
Ross Wightman c0d7388a1b Improving kwarg merging in more models
2 years ago
Ross Wightman 60ebb6cefa Re-order vit pretrained entries for more sensible default weights (no .tag specified)
2 years ago
Ross Wightman e861b74cf8 Pass through --model-kwargs (and --opt-kwargs for train) from command line through to model __init__. Update some models to improve arg overlay. Cleanup along the way.
2 years ago
Ross Wightman add3fb864e Working on improved model card template for push_to_hf_hub
2 years ago
Ross Wightman 6e5553da5f
Add ConvNeXt-V2 support (model additions and weights) (#1614)
2 years ago
Ross Wightman 6902c48a5f Fix ResNet based models to work w/ norm layers w/o affine params. Reformat long arg lists into vertical form.
2 years ago
Ross Wightman 8ece53e194 Switch BEiT to HF hub weights
2 years ago
Ross Wightman 9a51e4ea2e Add FlexiViT models and weights, refactoring, push more weights
2 years ago
Fredo Guan 10b3f696b4
Davit std (#6)
2 years ago
Ross Wightman 656e1776de Convert mobilenetv3 to multi-weight, tweak PretrainedCfg metadata
2 years ago