mradermacher/HOI-R1_Qwen2.5-VL-3B-Instruct-GGUF

Same weights, converted (fewer layers) vs thxplz/HOI-R1_Qwen2.5-VL-3B-Instruct. mradermacher/HOI-R1_Qwen2.5-VL-3B-Instruct-GGUF matches part of thxplz/HOI-R1_Qwen2.5-VL-3B-Instruct (435 of its 825 tensors; the rest of thxplz/HOI-R1_Qwen2.5-VL-3B-Instruct has no counterpart). On the shared part: Same weights stored in another dtype (BF16 → F16/F32).

Tensors changed: Embeddings 0/1, Norms 0/73, MLP 0/108, Attention 0/252, lm_head 0/1

On the Hub · Downloads on Model Pulse