rgtjf/Qwen2-UtK-ChatQA2-7B-128K

Full fine-tune vs rgtjf/Qwen2-UtK-7B-128K. Fine-tune: 93% of the weights changed, across 28 of 28 layers.

Tensors changed: Embeddings 0/1, Norms 55/57, MLP 84/84, Attention 196/196, lm_head 1/1

1 derivatives checked

• mradermacher/Qwen2-UtK-ChatQA2-7B-128K-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse