prithivMLmods/SAGE-MM-Qwen3-VL-4B-SFT_RL-GGUF

Same weights, converted (fewer layers) vs allenai/SAGE-MM-Qwen3-VL-4B-SFT_RL. prithivMLmods/SAGE-MM-Qwen3-VL-4B-SFT_RL-GGUF matches part of allenai/SAGE-MM-Qwen3-VL-4B-SFT_RL (399 of its 714 tensors; the rest of allenai/SAGE-MM-Qwen3-VL-4B-SFT_RL has no counterpart). On the shared part: Same weights stored in another dtype (F32 → F16/F32).

Tensors changed: Embeddings 0/1, Norms 0/145, MLP 0/108, Attention 0/144, lm_head 0/1

On the Hub · Downloads on Model Pulse