aesir-unlimited/AESIR-Hacker-Micro

Partial fine-tune vs Qwen/Qwen3.5-4B. Partial change: 19% of the weights changed (attention, mlp, norms), in 18 of 32 layers.

Tensors changed: Embeddings 0/5, MLP 32/199, Attention 13/324, Norms 3/209, other 0/1

1 derivatives checked

• mradermacher/AESIR-Hacker-Micro-GGUF: Same weights, converted

1 other repos hold byte-for-byte identical weight files.

On the Hub · Downloads on Model Pulse