Lambent/danube2-upscale-1

Partial fine-tune vs h2oai/h2o-danube2-1.8b-base. Partial change: 4% of the weights changed (attention, mlp, norms), in 1 of 24 layers.

Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 2/49, MLP 3/72, Attention 4/96

1 derivatives checked

• mradermacher/danube2-upscale-1-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse