quantized duplicated precisions.json, which is authoritative for the
quantized/non-quantized partition. default expressed a per-repository
recommendation nothing consumed: the library resolves its own per-platform
default. Both are gone from the schema.
Adds library_name to the model card, and pipeline_tag where it was missing.
Brings the config and card licenses into agreement, and repairs the variant
declarations that did not match the files actually published here.
Same rewrite as all-MiniLM-L6-v2: attention runs as [H, T, D] batched matmuls
instead of reshaping to 4-D, which removes the materialised view dispatches
those reshapes produce on Vulkan. Bit-identical on CPU.
fp16 is what makes this model worth lowering: at fp32 Vulkan loses to XNNPACK
on an Adreno 840 (0.91x), at fp16 it wins 1.35x, and on a Mali-G76 it is 1.90x.
Halving the file to 218 MB also eases the memory pressure that dominates this
model on 8 GB devices.
Mean pooling is written with keepdim so the reduction stays on the GPU, which
costs one delegate blob less. The 30527-row token embedding is kept off
texture storage, without which fp16 returns silently wrong values
(pytorch/executorch#22333); cosine against the CPU reference is 0.99995 on
Adreno and 0.99997 on Mali.
Requires the ExecuTorch 1.4.1 Vulkan runtime from the v0.10.0-libs-1.4.1
artifacts.
