namepd/Qwen3-VL-8B-Instruct-Q4_K_M-layers

Quantization (fewer layers) vs Qwen/Qwen3-VL-8B-Instruct-GGUF. namepd/Qwen3-VL-8B-Instruct-Q4_K_M-layers matches part of Qwen/Qwen3-VL-8B-Instruct-GGUF (11 of its 399 tensors; the rest of Qwen/Qwen3-VL-8B-Instruct-GGUF has no counterpart). On the shared part: namepd/Qwen3-VL-8B-Instruct-Q4_K_M-layers is a GGUF quantization (Q4_K/Q6_K) of the same weights: 4 tensors match exactly and 7 within quantization error.

Tensors changed: Norms 0/4, MLP 0/3, Attention 0/4

On the Hub · Downloads on Model Pulse