sqlite-vec/benchmarks-ann/datasets/nyt-768/distill-model.py

# /// script
# requires-python = ">=3.12"
# dependencies = [
#     "model2vec[distill]",
#     "torch<=2.7",
# ]
# ///

from model2vec.distill import distill

model = distill(model_name="BAAI/bge-base-en-v1.5", pca_dims=768)
model.save_pretrained("bge-base-en-v1.5-768")
print("Saved distilled model to bge-base-en-v1.5-768/")
Add comprehensive ANN benchmarking suite (#279) Extend benchmarks-ann/ with results database (SQLite with per-query detail and continuous writes), dataset subfolder organization, --subset-size and --warmup options. Supports systematic comparison across flat, rescore, IVF, and DiskANN index types. 2026-03-31 01:29:49 -07:00			`# /// script`
			`# requires-python = ">=3.12"`
			`# dependencies = [`
			`# "model2vec[distill]",`
			`# "torch<=2.7",`
			`# ]`
			`# ///`

			`from model2vec.distill import distill`

			`model = distill(model_name="BAAI/bge-base-en-v1.5", pca_dims=768)`
			`model.save_pretrained("bge-base-en-v1.5-768")`
			`print("Saved distilled model to bge-base-en-v1.5-768/")`