mradermacher/Qwen2.5-VL-7B-NSFW-Caption-V3-abliterated-GGUF

Same weights, converted (fewer layers) vs shutkit/Qwen2.5-VL-7B-NSFW-Caption-V3-abliterated. mradermacher/Qwen2.5-VL-7B-NSFW-Caption-V3-abliterated-GGUF matches part of shutkit/Qwen2.5-VL-7B-NSFW-Caption-V3-abliterated (339 of its 729 tensors; the rest of shutkit/Qwen2.5-VL-7B-NSFW-Caption-V3-abliterated has no counterpart). On the shared part: Same weights stored in another dtype (BF16 → F16/F32).

Tensors changed: Embeddings 0/1, Norms 0/57, MLP 0/84, Attention 0/196, lm_head 0/1

On the Hub · Downloads on Model Pulse