YannQi/R-4B

Full fine-tune (fewer layers) vs Qwen/Qwen3-4B. Qwen/Qwen3-4B matches part of YannQi/R-4B (397 of its 827 tensors; the rest of YannQi/R-4B has no counterpart). On the shared part: Fine-tune: 100% of the weights changed, across 36 of 36 layers.

Tensors changed: Norms 145/145, MLP 108/108, Attention 144/144

1 derivatives checked

• infil00p/R-4B-GGUF: Same weights, converted (fewer layers)

On the Hub · Downloads on Model Pulse