predibase/lorax Issues
Can't run Mistral quantized on T4
Updated 5log arbitrary headers
Updated 1Add support for LoReFT
Updated 1Upgrade to AWQ kernels v0.0.6
Updated 2Add support for fp8 (H100)
Updated 2Fix Mixtral usage with `--compile`
Updated 1Fix Mixtral with Medusa
ClosedNone
Closed 1Add helm chart to OCI repo
Updated 3Using Source = Local for Base Model
Updated 6Allow ignore unknown launcher args
Closed 3Want Lorax with newer version of TGI
Updated 5performance issue
Updated 4Invoke API based external models
Updated 1