nuetee/RA-RFT-Qwen2.5-VL-7B

Partial fine-tune vs Boshenxx/Time-R1-7B. Partial change: 86% of the weights changed (attention, lm_head, mlp, norms), in 28 of 32 layers.

Tensors changed: Embeddings 0/2, Norms 53/122, MLP 88/280, Attention 195/324, lm_head 1/1

2 derivatives checked

• mradermacher/RA-RFT-Qwen2.5-VL-7B-i1-GGUF: Quantization (fewer layers)
• mradermacher/RA-RFT-Qwen2.5-VL-7B-GGUF: Same weights, converted (fewer layers)

On the Hub · Downloads on Model Pulse