Buckets:
| # /// script | |
| # dependencies = [ | |
| # "transformers[serving] @ git+https://github.com/huggingface/transformers.git@main", | |
| # "torch", | |
| # "accelerate", | |
| # "requests", | |
| # "kernels>=0.11.0", | |
| # ] | |
| # /// | |
| import subprocess | |
| import sys | |
| MODEL = "mistralai/Mixtral-8x7B-Instruct-v0.1" | |
| PORT = 8000 | |
| cmd = [ | |
| "transformers", "serve", MODEL, | |
| "--continuous-batching", | |
| "--attn-implementation", "kernels-community/flash-attn2", | |
| "--dtype", "bfloat16", | |
| "--host", "0.0.0.0", | |
| "--port", str(PORT), | |
| ] | |
| print("Running:", " ".join(cmd), flush=True) | |
| sys.exit(subprocess.call(cmd)) | |
Xet Storage Details
- Size:
- 590 Bytes
- Xet hash:
- 1bc937ea08db246c47d26c57c4e501457d53696913a05bbe1f84b06e97999d9c
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.