DINOv2 'register' ViT model weights added (, ).DFN (Data Filtering Networks) and MetaCLIP ViT weights added.vision_transformer.py typing and doc cleanup by Laureηt.Updated imagenet eval and test set csv files with latest models.model_args will be passed as kwargs through to models on creation. Added significant flexibility for Hugging Face Hub based timm models via model_args config entry.Fix Python 3.7 compat, will be dropping support for it soon.Added EfficientViT-Large models, thanks SeeFun.Previous 0.6.x can be cloned from 0.6.x branch or installed via pip with version.Model cards include link to papers, original source, license. The Hugging Face Hub ( ) is now the primary source for timm weights.These can be passed to _checkpoint(., filter_fn=_transformer_v2.checkpoint_filter_fn) to remap your existing checkpoint. A number of models had their checkpoints remaped to match architecture changes needed to better support features_only=True, there are checkpoint_filter_fn methods in any model module that was remapped.There are deprecation mappings for these. In adding pretrained tags, many model names that existed to differentiate were renamed to use the tag (ex: vit_base_patch16_224_in21k -> vit_base_patch16_224.augreg_in21k).Using only architecture defaults to the first weights in the default_cfgs for that model architecture. The pretrained_tag is the specific weight variant (different head) for the architecture.All models now support architecture.pretrained_tag naming (ex resnet50.rsb_a1).Builder, helper, non-model modules in timm.models have a _ prefix added, ie -> timm.models._helpers, there are temporary deprecation mapping files but those will be removed.import or from import name needs to be changed now.from import name will still work via deprecation mapping (but please transition to timm.layers).They were previewed in 0.8.x dev releases but not everyone transitioned. Many changes since the last 0.6.x stable releases. ❗Updates after are available in version >= 0.9❗
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |