mergekit-community/Qwen2-2B-Dolphin-RepleteCoder

Partial fine-tune vs dphn/dolphin-2.9.3-qwen2-1.5b. Partial change: 49% of the weights changed (attention, mlp, norms), in 16 of 28 layers.

Tensors changed: Embeddings 0/1, Norms 33/57, MLP 48/84, Attention 112/196

1 derivatives checked

• mradermacher/Qwen2-2B-Dolphin-RepleteCoder-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse