ABX-AI/Infinitely-Kunodiculous-9B

Partial fine-tune vs ChaoticNeutrals/Infinitely-Laydiculous-9B. Partial change: 50% of the weights changed (attention, lm_head, mlp, norms), in 20 of 40 layers.

Tensors changed: MLP 60/120, Norms 41/81, Attention 80/160, Embeddings 0/1, lm_head 1/1

2 derivatives checked

• mradermacher/Infinitely-Kunodiculous-9B-i1-GGUF: Quantization
• mradermacher/Infinitely-Kunodiculous-9B-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse