Space requirements for LLM Models?

[goh__richard]

HI

How  much disk space is needed for the new LLM models like

Qwen, DeepSee?

Thank you

[Aznie_Intel]

Hi Goh_Richard,

Thanks for reaching out. The disk space needed for LLM models like Qwen and DeepSeek depends on the model size and precision. Here’s a quick overview:

Qwen Models:

• Qwen-7B: ~16.1 GB (FP16)
• Qwen-14B: ~32.2 GB (FP16)
• Qwen-72B: ~165.6 GB (FP16)

DeepSeek Models:

• DeepSeek-R1 (Full Model): ~1,342 GB (FP16)
• Distilled versions (e.g., DeepSeek-R1-Distill-Qwen-7B): ~16 GB

You may also reduce the storage needs by performing Model Optimizatrion -NNCF.

I hope this helps.

Regards,

Aznie

[goh__richard]

Thank you.

Not small , lol

[Aznie_Intel]

Hi Goh_Richard,

This thread will no longer be monitored since this issue has been resolved. If you need any additional information from Intel, please submit a new question. 

Regards,

Aznie