minicpm-llama3-2.5-8b-16-v With only 8B parameters, it surpasses widely used proprietary models like GPT-4V-1106, Gemini Pro, Claude 3 and Qwen-VL-Max and greatly outperforms other Llama 3-based MLLMs
377 Pulls Updated 5 months ago
073e5b8d024e · 103B
{
"num_gpu": 12,
"num_keep": 32,
"num_predict": -2,
"stop": [
"[<|eot_id|>"
],
"top_k": 33,
"top_p": 0.66
}