Interrupted model upload resume | Generated by AI
Here’s the status:
Analysis (done)
- Model: GPT-2 124M (12 layers, 768 hidden, 12 heads, 1024 ctx, vocab 50257)
- Files: config.json, generation_config.json, tokenizer.json, tokenizer_config.json, model.safetensors (622MB float32)
- Missing: README.md (model card) — fixed, uploaded to server
Upload progress
- Repo created: https://huggingface.co/lzwjava/sec-edgar-gpt-124m-hf ✓
- hf CLI installed on server (47.86.12.246) ✓
- HF token configured ✓
- README.md (model card) created + copied to server ✓
- Upload started — got 133MB/622MB into model.safetensors before interrupted (signal 15, likely SSH timeout)
What’s left
- Resume the upload. The small files (5/6) already transferred, only model.safetensors remains (~489MB left)
Want me to resume the upload? I’ll run it in background so it doesn’t timeout:
ssh root@47.86.12.246 "nohup hf upload lzwjava/sec-edgar-gpt-124m-hf ~/sec-edgar-gpt-124m-hf/ > /tmp/hf-upload.log 2>&1 &"
Or I can start it now with a longer timeout. Which do you prefer?