zjunlp/OceanGPT-basic-4B-Thinking

Full fine-tune vs Qwen/Qwen3-4B-Thinking-2507. Fine-tune: 100% of the weights changed, across 36 of 36 layers.

Tensors changed: Embeddings 1/1, Norms 137/145, MLP 108/108, Attention 144/144

2 derivatives checked

• mradermacher/OceanGPT-basic-4B-Thinking-i1-GGUF: Quantization
• mradermacher/OceanGPT-basic-4B-Thinking-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse