mradermacher/VST-7B-RL-GGUF

Same weights, converted (fewer layers) vs rayruiyang/VST-7B-RL. mradermacher/VST-7B-RL-GGUF matches part of rayruiyang/VST-7B-RL (339 of its 729 tensors; the rest of rayruiyang/VST-7B-RL has no counterpart). On the shared part: Same weights stored in another dtype (BF16 → F16/F32).

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

On the Hub · Downloads on Model Pulse