Etherll/Tashkeel-700M

Full fine-tune vs LiquidAI/LFM2-700M. Fine-tune: 100% of the weights changed, across 16 of 16 layers.

Tensors changed: Embeddings 1/2, other 30/30, MLP 48/48, Norms 23/44, Attention 24/24

2 derivatives checked

• mradermacher/Tashkeel-700M-i1-GGUF: Full fine-tune
• mradermacher/Tashkeel-700M-GGUF: Full fine-tune

On the Hub · Downloads on Model Pulse