História revízii

Autor SHA1 Správa Dátum
  Matthias Reso e2f77dbc21 fix quant config 9 mesiacov pred
  Matthias Reso 6ef9a78458 Fix issues with quantization_config == None 9 mesiacov pred
  Matthias Reso 0920b1a415 Fix quantization for inference 9 mesiacov pred
  Hamid Shojanazeri d51d2cce9c adding sdpa for flash attn 1 rok pred
  Hamid Shojanazeri db8af96ff0 update the model load with native flash attn 1 rok pred
  Matthias Reso 4c9cc7d223 Move modules into separate src folder 1 rok pred