Wiself/Holodeck-Lounge-MTP

Full fine-tune vs Jackrong/Qwopus3.5-9B-Coder. Fine-tune: 100% of the weights changed, across 32 of 32 layers.

Tensors changed: lm_head 1/1, Embeddings 5/5, MLP 211/211, Attention 336/336, other 0/1, Norms 221/221

2 derivatives checked

• mradermacher/Holodeck-Lounge-MTP-i1-GGUF: Quantization
• mradermacher/Holodeck-Lounge-MTP-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse