Enhancement: Verify model capabilities before downloading

[Mte90]

So I tried this parameters --model CodeLlama-7B --chat-model CodeGemma-7B, tabby downloaded and after that checked if the chat model has a chat_template.
Instead it should do that before starting downloading it.

Writing to new file.
🎯 Downloaded https://huggingface.co/TheBloke/CodeLlama-7B-GGUF/resolve/main/codellama-7b.Q4_0.gguf to /home/mte90/.tabby/models/TabbyML/CodeLlama-7B/ggml/model-00001-of-00001.gguf.tmp
   00:05:34 ▕████████████████████▏ 3.56 GiB/3.56 GiB  10.91 MiB/s  ETA 0s.                                                                                                                                         Writing to new file.
🎯 Transferring https://huggingface.co/bartowski/codegemma-7b-GGUF/resolve/main/codegemma-7b-Q4_K_M.gguf
🎯 Downloaded https://huggingface.co/bartowski/codegemma-7b-GGUF/resolve/main/codegemma-7b-Q4_K_M.gguf to /home/mte90/.tabby/models/TabbyML/CodeGemma-7B/ggml/model-00001-of-00001.gguf.tmp
⠼     5.269 s███Starting...The application panicked (crashed).B/s  ETA 0s.                                                                                                                                         
Message:  Chat model requires specifying prompt template
Location: /__w/tabby/tabby/crates/llama-cpp-server/src/lib.rs:225

Backtrace omitted. Run with RUST_BACKTRACE=1 environment variable to display it.
Run with RUST_BACKTRACE=full to include source snippets.