Repository Issue Activity (beta)

#2789 Speed up HuggingFace safetensor conversion via backend-native loading

#1004 Add preprocessor for t5 model

#1963 Masked BERT model from keras_nlp produces weird results

#2055 The attention scores are always `None` in `CachedMultiHeadAttention`

#1907 Add link for Bounding Box guides from keras-core.

#784 BertClassifier SavedModels don't include the pre-processor.

#973 Serialization issue with custom learning rate schedule

#2780 [Gemma3] Significant time spent on Concatenate op during KV Cache update

#2364 Add LTX-Video

#2701 LoRA save/load Bug

#1061 keras_nlp.models.OPTBackbone.from_preset fails with max_sequence_length=256

#1236 Is it possible to apply GPU Delegate to the transformer models?

#2783 fix: tar archive extraction vulnerable to symlink-based path traversal

#2128 Allow computer vision task models to run without tf.data

#2333 Add Devstral Small 1.1

#2349 Add Voxtral

#2470 Implement Llama 3.2 Vision (Multimodal) Architecture

#989 Maybe throw an error in `generate` if inputs is dense token ids and `max_length` is set

#797 Incorrect Detokenisation for Special Tokens by SentencePiece-based Tokenizers

#1312 Remove workaround for `take_along_axis`

#1479 How to add a serialized model and weights of a keras model to keras-nlp?

#2101 KerasHub has mandatory dependency on tensorflow-text but it isn't mandatory

#2669 CONTRIBUTING_MODELS.md is outdated

#1988 Missing Transformers initializer for Falcon models

#1985 tf2.16.1 isn't compatibale with tensorflow text 2.10