REF: Remove some builtin old models and `ggmlv3` model format

[ChengjieLi28]

1. Remove some builtin old models

• Baichuan baichuan-chat
• Starcoder starcoderplus starchat
• glaive-coder
• wizardlm-v1.0
• vicuna-v1.3 vicuna-v1.5
• OpenBuddy
• orca
• falcon
• Chatglm chatglm-2
• Tiny-llama (TODO)
• opt (TODO)

2. Change llama-2 / llama-2-chat ggmlv3 format to ggufv2 format.
3. Remove support for ggmlv3 format.
4. Remove register model s3 schema support.
5. Remove support for self-hosted models.
6. Remove some unused codes.
7. Remove opencv for a required dependency. Place it in all dependency.
8. Rename pytorch dir to transformers and rename ggml dir to llama_cpp.