mradermacher/Behemoth-3B-070225-post0.1-GGUF

Same weights, converted (fewer layers) vs prithivMLmods/Behemoth-3B-070225-post0.1. mradermacher/Behemoth-3B-070225-post0.1-GGUF matches part of prithivMLmods/Behemoth-3B-070225-post0.1 (434 of its 824 tensors; the rest of prithivMLmods/Behemoth-3B-070225-post0.1 has no counterpart). On the shared part: Same weights stored in another dtype (BF16/F16 → F16/F32).

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

On the Hub · Downloads on Model Pulse