Delete converted_tokenizer

Delete converted_tokenizer_128k

Upload 8 files

Create phi_convert_hf_gguf

Update convert_tokenizer.py

Update convert_tokenizer.py

Update convert_tokenizer.py

Update README.md

Update README.md

tokenizer trained on 128k - not sure if it'll be better for RoPE or not

Update README.md

Update README.md

Upload 9 files

Update LICENSE

initial commit