[New Model]: Add support for Ovis models

[ismael-dm]

🚀 The feature, motivation and pitch

Hi, I wanted to try the model AIDC-AI/Ovis1.6-Gemma2-9B but I´m getting the error "Model architectures ['Ovis'] are not supported for now". Is it planned for the future or is there some workaround?

Alternatives

No response

Additional context

The full error that I´m getting is:

ValueError: Model architectures ['Ovis'] are not supported for now. Supported architectures: ['AquilaModel', 'AquilaForCausalLM', 'BaiChuanForCausalLM', 'BaichuanForCausalLM', 'BloomForCausalLM', 'ChatGLMM odel', 'ChatGLMForConditionalGeneration', 'CohereForCausalLM', 'DbrxForCausalLM', 'DeciLMForCausalLM', 'DeepseekForCausalLM', 'DeepseekV2ForCausalLM', 'ExaoneForCausalLM', 'FalconForCausalLM', 'GemmaForCau salLM', 'Gemma2ForCausalLM', 'GPT2LMHeadModel', 'GPTBigCodeForCausalLM', 'GPTJForCausalLM', 'GPTNeoXForCausalLM', 'InternLMForCausalLM', 'InternLM2ForCausalLM', 'JAISLMHeadModel', 'LlamaForCausalLM', 'LLaM AForCausalLM', 'MistralForCausalLM', 'MixtralForCausalLM', 'QuantMixtralForCausalLM', 'MptForCausalLM', 'MPTForCausalLM', 'MiniCPMForCausalLM', 'MiniCPM3ForCausalLM', 'NemotronForCausalLM', 'OlmoForCausalL M', 'OlmoeForCausalLM', 'OPTForCausalLM', 'OrionForCausalLM', 'PersimmonForCausalLM', 'PhiForCausalLM', 'Phi3ForCausalLM', 'PhiMoEForCausalLM', 'Qwen2ForCausalLM', 'Qwen2MoeForCausalLM', 'Qwen2VLForConditi onalGeneration', 'RWForCausalLM', 'StableLMEpochForCausalLM', 'StableLmForCausalLM', 'Starcoder2ForCausalLM', 'SolarForCausalLM', 'ArcticForCausalLM', 'XverseForCausalLM', 'Phi3SmallForCausalLM', 'MedusaMo del', 'EAGLEModel', 'MLPSpeculatorPreTrainedModel', 'JambaForCausalLM', 'GraniteForCausalLM', 'MistralModel', 'Blip2ForConditionalGeneration', 'ChameleonForConditionalGeneration', 'FuyuForCausalLM', 'Inter nVLChatModel', 'LlavaForConditionalGeneration', 'LlavaNextForConditionalGeneration', 'LlavaNextVideoForConditionalGeneration', 'LlavaOnevisionForConditionalGeneration', 'MiniCPMV', 'PaliGemmaForConditional Generation', 'Phi3VForCausalLM', 'PixtralForConditionalGeneration', 'QWenLMHeadModel', 'UltravoxModel', 'MllamaForConditionalGeneration', 'BartModel', 'BartForConditionalGeneration']

Before submitting a new issue...

• [X] Make sure you already searched for relevant issues, and asked the chatbot living at the bottom right corner of the documentation page, which can answer lots of frequently asked questions.

[DarkLight1337]

Yeah, it's not implemented in vLLM yet. Any help is welcome!