Hcompany/Holo1.5-72B

Full fine-tune vs Qwen/Qwen2.5-VL-72B-Instruct. Fine-tune: 100% of the weights changed, across 80 of 80 layers.

Tensors changed: Embeddings 2/2, Norms 145/226, Attention 688/688, MLP 436/436, lm_head 1/1

2 derivatives checked

• mradermacher/Holo1.5-72B-i1-GGUF: Quantization
• mradermacher/Holo1.5-72B-GGUF: Quantization

On the Hub · Downloads on Model Pulse