Melikshah's picture
download
raw
1.44 kB
import os
import sys
from pathlib import Path
_EARLY = [m for m in ("huggingface_hub", "transformers", "mlx_lm", "matplotlib") if m in sys.modules]
if _EARLY:
raise ImportError(f"import opharm before {', '.join(_EARLY)} so every cache stays inside the project")
ROOT = Path(__file__).resolve().parents[2]
CACHE = ROOT / ".cache"
CONFIGS = ROOT / "configs"
DATA = ROOT / "data"
BENCH = ROOT / "benchmark"
RUNS = ROOT / "runs"
RESULTS = ROOT / "results"
_LOCAL = {
"HF_HOME": CACHE / "huggingface",
"HF_HUB_CACHE": CACHE / "huggingface" / "hub",
"HF_XET_CACHE": CACHE / "huggingface" / "xet",
"TORCH_HOME": CACHE / "torch",
"MPLCONFIGDIR": CACHE / "matplotlib",
"XDG_CACHE_HOME": CACHE,
}
for _k, _v in _LOCAL.items():
os.environ[_k] = str(_v)
os.environ.setdefault("HF_HUB_DISABLE_TELEMETRY", "1")
os.environ.setdefault("TOKENIZERS_PARALLELISM", "false")
os.environ.setdefault("HF_DEACTIVATE_ASYNC_LOAD", "1")
def _dotenv(path):
out = {}
if path.is_file():
for line in path.read_text().splitlines():
line = line.strip()
if line and not line.startswith("#") and "=" in line:
key, value = line.removeprefix("export ").split("=", 1)
out[key.strip()] = value.strip().strip("'\"")
return out
if not os.environ.get("HF_TOKEN"):
_token = _dotenv(ROOT / ".env").get("HF_READ")
if _token:
os.environ["HF_TOKEN"] = _token

Xet Storage Details

Size:
1.44 kB
·
Xet hash:
53fe37e3aaf188045dc779d5ae529be11aacdc6b703fcd234a5bdec9ecd6689e

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.