eaddario/Apriel-1.6-15b-Thinker-GGUF

Same weights, converted (fewer layers) vs ServiceNow-AI/Apriel-1.6-15b-Thinker. eaddario/Apriel-1.6-15b-Thinker-GGUF matches part of ServiceNow-AI/Apriel-1.6-15b-Thinker (435 of its 657 tensors; the rest of ServiceNow-AI/Apriel-1.6-15b-Thinker has no counterpart). On the shared part: Same weights stored in another dtype (BF16 → F16/F32).

Tensors changed: Embeddings 0/1, Norms 0/97, MLP 0/144, Attention 0/192, lm_head 0/1

On the Hub · Downloads on Model Pulse