luodi-7
/

internvl_memeloc_stage2

Model card Files Files and versions Community

YAML Metadata Warning: empty or missing yaml metadata in repo card (https://huggingface.co/docs/hub/model-cards#model-card-metadata)

模型介绍

微调internvl2_5-4B
希望能微调模型根据一串meme文字和meme底图学会输出【文字框位置+对应位置填的文字】
在一阶段的基础上，mask掉文字部分的loss,只计算预测框的loss,再进行训练希望提高预测框准确率

测试结果

avg_iou=0.4272
avg_similarity=0.9772

Downloads last month: 5

Safetensors

Model size

3.71B params

Tensor type

BF16

·

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support