Koopah/Qwen3.6-35B-A3B-NVFP4-DSPARK-v2-GGUF

Same weights, converted vs Koopah/Qwen3.6-35B-A3B-NVFP4-DSPARK-v2. Same weights stored in another dtype (BF16 → F16/F32), plus 3 new tensors.

Tensors changed: other 0/3, Norms 0/25, MLP 0/18, Attention 0/24

On the Hub · Downloads on Model Pulse