Add a missing step to the gpt4all instructions

[ThatcherC]

migrate-ggml-2023-03-30-pr613.py is needed to get gpt4all running.

When following the README instructions for gpt4all, I encounted the following error:

data/llama.cpp$ python3 convert-gpt4all-to-ggml.py models/gpt4all-models/gpt4all-lora-quantized.bin models/tokenizer.model

/data/llama.cpp$ ./main -m models/gpt4all-models/gpt4all-lora-unfiltered-quantized.bin -n 128
main: seed = 1680382134
llama_model_load: loading model from 'models/gpt4all-models/gpt4all-lora-unfiltered-quantized.bin' - please wait ...
models/gpt4all-models/gpt4all-lora-unfiltered-quantized.bin: invalid model file (bad magic [got 0x67676d66 want 0x67676a74])
    you most likely need to regenerate your ggml files
    the benefit is you'll get 10-100x faster load times
    see https://github.com/ggerganov/llama.cpp/issues/91
    use convert-pth-to-ggml.py to regenerate from original pth
    use migrate-ggml-2023-03-30-pr613.py if you deleted originals
llama_init_from_file: failed to load model
main: error: failed to load model 'models/gpt4all-models/gpt4all-lora-unfiltered-quantized.bin'

I think the gpt4all conversion script is a bit out of date and produces files that need to be converted with migrate-ggml-2023-03-30-pr613.py.

I was able to run python3 migrate-ggml-2023-03-30-pr613.py models/gpt4all-models/gpt4all-lora-unfiltered-quantized.bin models/gpt4all-models/gpt4all-lora-unfiltered-quantized-pr613.bin which seemed to do the trick. After that, ./main -m models/gpt4all-models/gpt4all-lora-unfiltered-quantized-pr613.bin ran successfully!

This PR updates the README to add that migrate-ggml-2023-03-30-pr613.py step.

[sw]

Yes, the *-to-ggml.py scripts need to be updated, I've opened an issue for this: #704

In the meantime, I think it makes sense to update the README.

Edit: looks like #545 might make this irrelevant, let's see if that one can be merged soon.