50h100a 273c61d406 guard against nan temperature from dynatemp (or anywhere else). vor 2 Monaten
..
guided_decoding 0256ed236b feat: windows support (#790) vor 2 Monaten
layers 273c61d406 guard against nan temperature from dynatemp (or anywhere else). vor 2 Monaten
model_loader 0f1af04cf5 frontend: minor logging improvements (#787) vor 2 Monaten
models 0256ed236b feat: windows support (#790) vor 2 Monaten
__init__.py 7df7b8ca53 optimization: reduce end-to-end overhead from python obj allocation (#666) vor 4 Monaten
_custom_op.py 5d37ec1016 suppress tpu import warning (#696) vor 4 Monaten
parameter.py f98e7b2f8c feat: add HQQ quantization support (#795) vor 2 Monaten
pooling_metadata.py f1d0b77c92 [0.6.0] Release Candidate (#481) vor 4 Monaten
sampling_metadata.py ed87b5dd32 remove special-case values for prompt sequences vor 4 Monaten
utils.py 9d81716bfd [v0.5.3] Release Candidate (#388) vor 8 Monaten