Spaces:
Running
on
CPU Upgrade
Some models have missing "Param# (B)".
For example all the base Llama models.
Hope the missing data can be fixed and displayed.
HI
@felixz
!
We can only extract the number of parameters if either 1) the model weights are saved using safetensors or 2) there is a parameter number in the name.
If you want the precise information for a specific model, please ask the authors to convert their weights on the hub to safetensors 🤗
HI @felixz !
We can only extract the number of parameters if either 1) the model weights are saved usingsafetensorsor 2) there is a parameter number in the name.
If you want the precise information for a specific model, please ask the authors to convert their weights on the hub tosafetensors🤗
So at least the Llama pretrained should be filled out:
Hi! Yes, that's funny, I'll have a look
Fixed, thank you for pointing it out!