Capacity-first experimental checkpoint: every Q4-family build in this sweep failed the predeclared minimum aligned-cosine fidelity gate. It is published as an experimental comparator, not as a BF16-equivalent embedding model.

This is the Q uniform affine quantization checkpoint from a matched local embedding-quantization experiment. It is published with explicit lineage, calibration evidence where applicable, and the bounded evaluation result that accompanied the conversion.

Provenance and lineage

Upstream model: Alibaba-NLP/gte-Qwen2-1.5B-instruct
Upstream revision recorded for publication: a9af15a6372d7d6b25e9fb07c2ccb9e1fe645644
Revision evidence: upstream revision verified at publication time; historical local snapshot metadata was not retained
Direct parent: TiGa-RCE/gte-Qwen2-1.5B-instruct-MLX-BF16
Conversion rule: every quantized checkpoint branches directly from the family MLX BF16 checkpoint; no lossy checkpoint was used to create another.
Quantization: Q uniform affine quantization, nominal 4-bit, group size 64
Local conversion stack: oMLX 0.5.3, mlx-lm 0.31.3, MLX 0.32.0
Full collection: MLX Embedding Quantization Matrix

PROVENANCE.json contains machine-readable lineage and SHA-256 hashes for the published weight files. No importance matrix was used for this checkpoint.

Bounded local evaluation

Metric	Result
Top-1 retrieval	1.000
MRR	1.000
Mean aligned cosine vs BF16	0.969969
Minimum aligned cosine vs BF16	0.948780
Score RMSE vs BF16	0.020860
Queries with rank change	0
Predeclared gate	FAIL

The evaluation used 24 frozen query/document pairs, the upstream query instruction recipe, last-token pooling, L2 normalization, and direct comparison with vectors from the family BF16 checkpoint. This is an engineering smoke test, not MTEB and not a claim of universal quality. Retrieval success and representation fidelity are reported separately.

Runtime scope

This checkpoint targets Apple Silicon through MLX/oMLX. CUDA and PyTorch results are a separate control lane and must not be interpreted as measurements of MLX/Metal kernel performance.

License and attribution

Apache-2.0, following the upstream model card. The original model authors retain attribution for the upstream model; this repository contains a local MLX conversion or quantized derivative prepared by TiGa-RCE for reproducibility research.

TiGa-RCE/gte-Qwen2-1.5B-instruct-MLX-Q4

gte-Qwen2-1.5B-instruct — MLX Q4

Provenance and lineage

Bounded local evaluation

Runtime scope

License and attribution

TiGa-RCE/
gte-Qwen2-1.5B-instruct-MLX-Q4