Kodjaoglanian/Qwen-3.5-4B-Auto-ASVD-GGUF

Same weights, converted vs Kodjaoglanian/Qwen-3.5-4B-Auto-ASVD. Same weights stored in another dtype (BF16 → F16/F32), plus 67 new tensors.

Tensors changed: Embeddings 0/1, Norms 0/81, MLP 0/75, Attention 0/10

On the Hub · Downloads on Model Pulse