| Models | Example Huggingface Models | Template | Modality | Model Size | Alignment Algorithm |
|---|---|---|---|---|---|
| LLaVA 1.5 | liuhaotian/llava-v1.5-7b liuhaotian/llava-v1.5-13b | llava | Image / Text to Text | 7B / 13B | SFT / PPO |
| LLaVA 1.6 (LLaVA-NeXT) | liuhaotian/llava-v1.6-vicuna-7b , liuhaotian/llava-v1.6-vicuna-13b , liuhaotian/llava-v1.6-34b | llava-next | Image / Video / Text to Text | 7B / 13B / 34B | SFT / PPO |
| InternVL | OpenGVLab/InternVL-14B-224px | internvl | Image / Video / Text to Text | 14B | SFT / PPO |
| InternVL2 | OpenGVLab/InternVL2-26B , OpenGVLab/InternVL2-8B , OpenGVLab/InternVL2-4B , OpenGVLab/InternVL2-2B | internvl2 | Image / Video / Text to Text | 2B / 4B / 8B /26B | SFT / PPO |
| Gemma-2 | google/gemma-2-9b , google/gemma-2-27b | gemma2 | Text to Text | 9B / 27B | SFT / PPO / DPO / SimPO / KTO / ORPO (All) |
| Gemma-1.1 / Gemma | google/gemma-1.1-2b-it , google/gemma-1.1-7b-it | gemma | Text to Text | 2B / 7B | All |
| Qwen2 | Qwen/Qwen2-0.5B , Qwen/Qwen2-1.5B , Qwen/Qwen2-7B , Qwen/Qwen2-72B | qwen2 | Text to Text | 0.5B / 1.5B / 7B / 72B | All |
| Qwen1.5 | Qwen/Qwen1.5-110B , Qwen/Qwen1.5-72B , Qwen/Qwen1.5-32B , Qwen/Qwen1.5-14B , Qwen/Qwen1.5-7B , Qwen/Qwen1.5-4B , Qwen/Qwen1.5-1.8B , Qwen/Qwen1.5-0.5B | qwen | Text to Text | 0.5B / 1.5B / 7B / 14B / 32B / 72B / 100B | All |
| Mixtral | mistralai/Mixtral-8x7B-Instruct-v0.1 | mixtral | Text to Text | 7B / 8X7B / 8X22B | All |
| Phi-3 | microsoft/Phi-3-mini-4k-instruct | phi3 | Text to Text | 3.8B | All |
| Llama3 | meta-llama/Meta-Llama-3-8B , meta-llama/Meta-Llama-3-70B | llama3 / llama3-nosys | Text to Text | 8B / 70B | All |
| Llama2 | meta-llama/Llama-2-7b-hf , meta-llama/Llama-2-13b-hf , meta-llama/Llama-2-70b-hf | llama2 / llama2-nosys | Text to Text | 7B / 13B / 70B | All |
| GLM-4 | THUDM/glm-4-9b | chatglm4 | Text to Text | 9B | All |
| ChatGLM3 | THUDM/chatglm3-6b | chatglm3 | Text to Text | 6B | All |
| DeepSeek-VL | deepseek-ai/deepseek-vl-1.3b-base , deepseek-ai/deepseek-vl-7b-base | deepseek-vl | Image / Text to Text | 1.3B / 7B | All |
| Baichuan 2 | baichuan-inc/Baichuan2-13B-Base , baichuan-inc/Baichuan2-7B-Base | baichuan2 | Text to Text | 7B / 13B | All |
| Vicuna | lmsys/vicuna-7b-v1.5 , lmsys/vicuna-13b-v1.5 , lmsys/vicuna-33b-v1.3 | vicuna | Text to Text | 7B / 13B / 33B | All |
| Alpaca | tatsu-lab/alpaca | alpaca | Text to Text | 7B | All |
| Dolphin | cognitivecomputations/dolphin-2.9-llama3-8b , cognitivecomputations/dolphin-2.5-mixtral-8x7b , cognitivecomputations/dolphin-2.9-mixtral-8x22b | dolphin | Text to Text | 8B / 8X7B / 8X22B | All |
| Qwen2-VL | Qwen/Qwen2-VL-7B-Instruct | Qwen2-VL | Image / Video / Text to Text | 7B | All |
| Qwen2-Audio | Qwen/Qwen2-Audio-7B-Instruct | Qwen2Audio | Audio / Text to Text | 7B | All |
| Chameleon | PKU-Alignment/AA-chameleon-7b-base, facebook/chameleon-7b | Chameleon | Image Text to Image Text | 7B | All |