[WIP] TextGrad Vision

[mertyg]

TextGrad vision integration! Still WIP.

• Let's add a better example notebook
• Let's add more docs
• Let's zoom into image types and whether there are failure cases @vinid
• Are we happy with how we are handling multimodal vs not? It seems brittle for now, there must be a better way than the current

def generate(self, content: Union[str, List[Union[str, bytes]]], system_prompt=None, **kwargs):
        if isinstance(content, str):
            return self._generate_text(content, system_prompt=system_prompt, **kwargs)

        elif isinstance(content, list):
            if (not self.is_multimodal):
                raise NotImplementedError("Multimodal generation is only supported for GPT-4 models.")

            return self._generate_multimodal(content, system_prompt=system_prompt, **kwargs)