etiennebamas/qwen3-sft-classic-small-data-equal-compute

Full fine-tune vs formalmathatepfl/qwen3-cpt. Fine-tune: 92% of the weights changed, across 36 of 36 layers.

Tensors changed: Embeddings 0/1, Norms 139/145, MLP 108/108, Attention 144/144, lm_head 1/1

On the Hub · Downloads on Model Pulse