Core ML weights in float32, with sequence length fixed to 64.
float32
64
Would it be possible to provide the script to convert the model into coreml? Wanted to try and convert falcon-rw-1b to coreml as well.
falcon-rw-1b
· Sign up or log in to comment
