NewEden/Apertus-8b-instruct-patched

Full fine-tune vs swiss-ai/Apertus-8B-2509. Fine-tune: 100% of the weights changed, across 32 of 32 layers.

Tensors changed: Embeddings 1/1, Norms 117/129, MLP 84/192, Attention 128/128, lm_head 1/1

2 derivatives checked

• mradermacher/Apertus-8b-instruct-patched-i1-GGUF: Quantization
• mradermacher/Apertus-8b-instruct-patched-GGUF: Same weights, converted

3 other repos hold byte-for-byte identical weight files.

On the Hub · Downloads on Model Pulse