export HF_HUB_ENABLE_HF_TRANSFER=1
huggingface-cli download Qwen/QwQ-32B-Preview
model-00007-of-00017.safetensors: 23%|█████████████████ | 912M/3.90G [00:23<00:45, 65.8MB/s]
model-00007-of-00017.safetensors: 99%|████████████████████████████████████████████████████████████████████████▌| 3.88G/3.90G [05:45<01:32, 228kB/s
This only happens in some clusters.
Could anyone help me figure out the possible cause?