cerebras/Kimi-Linear-REAP-35B-A3B-Instruct
LoRA-style vs moonshotai/Kimi-Linear-48B-A3B-Instruct. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of w1, w2, w3 changed, in 26 of 27 layers; norms, biases and embeddings are untouched.
Tensors changed: Attention 0/308, Embeddings 0/1, Norms 0/82, MLP 1717/1819, lm_head 0/1
3 derivatives checked
- mradermacher/Kimi-Linear-REAP-35B-A3B-Instruct-i1-GGUF: Quantization
- mradermacher/Kimi-Linear-REAP-35B-A3B-Instruct-GGUF: Quantization
- nightmedia/Kimi-Linear-REAP-35B-A3B-Instruct-mxfp4-mlx: Quantization