qihoo360/Light-R1-32B-DS

Full fine-tune vs deepseek-ai/DeepSeek-R1-Distill-Qwen-32B. Fine-tune: 98% of the weights changed, across 64 of 64 layers.

Tensors changed: Embeddings 0/1, Attention 448/448, MLP 192/192, Norms 126/129, lm_head 1/1

3 derivatives checked

• mradermacher/Light-R1-32B-DS-i1-GGUF: Quantization
• mradermacher/Light-R1-32B-DS-GGUF: Quantization
• tensorblock/Light-R1-32B-DS-GGUF: Quantization

1 other repos hold byte-for-byte identical weight files.

On the Hub · Downloads on Model Pulse