NoemaAI-labs/Noema-Overfit

Quantization vs Qwen/Qwen3.5-122B-A10B. NoemaAI-labs/Noema-Overfit is a GGUF quantization (Q4_K/Q5_K) of the same weights: 169 tensors match exactly and 194 within quantization error.

Tensors changed: lm_head 0/1, Embeddings 0/1, Attention 0/48, MLP 0/192, Norms 0/121

On the Hub · Downloads on Model Pulse