Hcompany/Holo-3.1-0.8B

Full fine-tune vs Qwen/Qwen3.5-0.8B. Fine-tune: 100% of the weights changed, across 24 of 24 layers.

Tensors changed: Attention 199/216, Norms 108/129, Embeddings 4/4, MLP 122/124

3 derivatives checked

• mradermacher/Holo-3.1-0.8B-i1-GGUF: Quantization
• mradermacher/Holo-3.1-0.8B-GGUF: Same weights, converted
• usermma/Holo-3.1-0.8B-abliterated-rdo: Partial fine-tune (fewer layers)

On the Hub · Downloads on Model Pulse