All models

DeepSeek-VL2

byDeepSeekDeepSeek· 13 Dec 2024
General purposeMultimodal

DeepSeek has released a new version of their vision models. The models come in various small sizes and are all MoE models, which is unusual in this size range. The language models are trained solely for this model, which could be from experimentation when training DeepSeek V3.

Specs
Params1.0B, 2.8B, 4.5B
LicenseDeepSeek Model License
Tags
modelsmodels/llmsmodels/llms/multi-modalartifacts/6
Adoption · Interconnects
RAM score
Hugging Face Downloads
5.2K
last 30d
179.5K
all time
HF Likes
387

Relative Adoption Metric contextualizes downloads against the model's size bucket.

Related Models