Convert Microsoft phi 1.5

[loretoparisi]

I'm trying to convert to GGUF the new phi-1_5

python3 convert.py /Downloads/phi-1_5 
Loading model file /Downloads/phi-1_5/pytorch_model.bin
Traceback (most recent call last):
  File "/llama.cpp/convert.py", line 1196, in <module>
    main()
  File "/llama.cpp/convert.py", line 1145, in main
    params = Params.load(model_plus)
             ^^^^^^^^^^^^^^^^^^^^^^^
  File "/llama.cpp/convert.py", line 299, in load
    params = Params.loadHFTransformerJson(model_plus.model, hf_config_path)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/llama.cpp/convert.py", line 209, in loadHFTransformerJson
    n_embd           = config["hidden_size"]
                       ~~~~~~^^^^^^^^^^^^^^^
KeyError: 'hidden_size'

Model Card is

• Architecture: Transformer-based model with next-word prediction objective
• Dataset size: 30B tokens
• Training tokens: 150B tokens
• Precision: fp16
• GPUs: 32xA100-40G
• Training time: 8 days

[ianscrivener]

Phi 1.5 is not (yet) a supported model README.md.

There is already an request in to add Phi 1.5 support [3139].