dhanesh-hf/Jarvis-Titan-M4-Activated

Partial fine-tune vs dhanesh-hf/Jarvis-Titan-V15-MoE-Decoupled. Partial change: 1% of the weights changed (attention, other), in 0 of 28 layers.

Tensors changed: Embeddings 0/1, Norms 0/71, MLP 0/196, Attention 21/217, other 42/63, lm_head 0/1

On the Hub · Downloads on Model Pulse