rmihaylov/falcontune Issues
FalconLite support
Updated 1Inference not accurate
Updatedmulti-gpu lora finetune falcon
Updated 108-bit models
Updatedfinetune on 4096 context length
Updated 2Possible to offload to cpu (ram)
Updated 5Metal support
UpdatedSupport multiturn chat ?
UpdatedUse prompt get error
Updated