deepseek-ai/DeepSeek-Prover-V1.5-SFT

Full fine-tune vs deepseek-ai/DeepSeek-Prover-V1.5-Base. Fine-tune: 100% of the weights changed, across 30 of 30 layers.

Tensors changed: Embeddings 1/1, Norms 61/61, lm_head 1/1, Attention 120/120, MLP 90/90

4 derivatives checked

• deepseek-ai/DeepSeek-Prover-V1.5-RL: Partial fine-tune
• mradermacher/DeepSeek-Prover-V1.5-SFT-GGUF: Same weights, converted
• Slim205/Lean_prover_v1: Full fine-tune
• Slim205/Lean-conjecturer: Full fine-tune

On the Hub · Downloads on Model Pulse