keitokei1994/Qwen3.5-2B-JP-v0.2

Partial fine-tune vs Qwen/Qwen3.5-2B. Partial change: 85% of the weights changed (attention, embeddings, mlp, norms), in 24 of 24 layers.

Tensors changed: Attention 134/264, Norms 46/177, Embeddings 1/4, MLP 72/172

2 derivatives checked

• mradermacher/Qwen3.5-2B-JP-v0.2-GGUF: Same weights, converted
• keitokei1994/Qwen3.5-2B-JP-v0.2-gguf: Quantization

On the Hub · Downloads on Model Pulse