Buckets:
| set -euo pipefail | |
| DATASET="YOUR_ORG/YOUR_DATASET" | |
| hf jobs run \ | |
| --name opensora-v21-reference-adapter \ | |
| --flavor a100-large \ | |
| --timeout 24h \ | |
| -s HF_TOKEN \ | |
| -v hf://creuthecat/Open-Sora-v2.1:/workspace/training:ro \ | |
| -v hf://openai/clip-vit-large-patch14:/clip:ro \ | |
| -v hf://datasets/${DATASET}:/data:ro \ | |
| -v hf://buckets/creuthecat/opensora-v21:/training-outputs \ | |
| nvcr.io/nvidia/pytorch:25.02-py3 \ | |
| bash -lc ' | |
| apt-get update && apt-get install -y ffmpeg git && | |
| python -m pip install transformers torchvision pillow safetensors && | |
| python /workspace/training/reference/train_reference_encoder.py \ | |
| --csv /data/train.csv \ | |
| --video-root /data/videos \ | |
| --output /training-outputs/reference_encoder/reference_encoder.safetensors \ | |
| --steps 10000 \ | |
| --lr 1e-5 | |
| ' | |
Xet Storage Details
- Size:
- 838 Bytes
- Xet hash:
- 6e8401e4cdcbca108d462897d38692b010a26d8c1ef0a34904390f535b5be398
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.