attn-signs/GPTR-8b-base

Full fine-tune vs yandex/YandexGPT-5-Lite-8B-pretrain. Fine-tune: 100% of the weights changed, across 32 of 32 layers.

Tensors changed: Norms 31/65, MLP 96/96, Attention 128/128

3 derivatives checked

• mradermacher/GPTR-8-base-i1-GGUF: Quantization
• mradermacher/GPTR-8-base-GGUF: Same weights, converted
• attn-signs/GPTR-8b-v1: Full fine-tune

On the Hub · Downloads on Model Pulse