contriever-mlx / README.md
masahiroid's picture
Publish validated original-checkpoint conversion with provenance and usage
618a741 verified
|
Raw History Blame Contribute Delete
3.51 kB
metadata
license: cc-by-nc-4.0
base_model: facebook/contriever
tags:
  - mlx
  - checkpoint-conversion

contriever — MLX

Dense retrieval embeddings with mask-aware mean pooling. Embeddings are not L2-normalized.

Conversion of facebook/contriever at revision 2bd46a25019aeea091fd42d1f0fd4801675cf699. No fine-tuning or replacement checkpoint. F32 baseline, not a 4-bit quantization. Original authors retain credit and license rights.

Usage and compatibility

Native MLX implementation with the source inference-model tensor names. This custom runtime is not an mlx-lm text-generation model.

Apple Silicon Mac required by the included MLX runtime. Inputs may vary in length (up to the source limit); compiled Apple exports below use fixed shapes.

pip install huggingface_hub mlx numpy transformers sentencepiece
from huggingface_hub import snapshot_download
from pathlib import Path
import sys, numpy as np
path = snapshot_download("masahiroid/contriever-mlx")
sys.path.insert(0, path)
from runtime_mlx import Model
model = Model(path)
from transformers import AutoTokenizer
tok = AutoTokenizer.from_pretrained(path, use_fast=True)
inputs = tok("Thank you so much!", return_tensors="np", truncation=True, max_length=128)
output = model(input_ids=inputs["input_ids"], attention_mask=inputs["attention_mask"])
print(np.array(output).shape)

Numerical validation

Compared against the original PyTorch implementation, 6 deterministic cases. Maximum absolute output difference: 1.2721866e-06; minimum cosine similarity: 0.99999993. See validation.json for each case. These are conversion parity checks, not a full task-quality benchmark or latency benchmark.

Text tests cover English emotions, neutral retrieval text, Japanese, empty input and truncation. Chronos tests cover a trend plus sinusoid, a constant series and a series with missing observations. Only applicable tests are used for each model.

Core ML/Core AI text exports use batch 1 and sequence length 128; tokenization must pad/truncate to 128. Chronos Apple exports use batch 1 and context length 2048. MLX/custom GGUF preserve the original maximum context rather than adding training. Short Chronos inputs can be left-padded with NaN for the fixed Apple exports. Forecast horizon remains 64.

Reproducibility

Source revision and output semantics: provenance.json. Conversion scripts are included. Reproduce in a checkout with sources/<original-model-name>/, sources.json and publish/ directories as in the scripts. PyTorch 2.14.1; Transformers 4.57.3; MLX 0.32.4; NumPy 1.26.4 (MLX/Core ML environment); coremltools 9.0; chronos-forecasting 2.3.2; GGUF 0.19.0; Core AI export uses coreai-torch 0.4.3. Core AI validation uses its separate Python 3.14 environment. Version compatibility outside the tested environment is not guaranteed.

License

Weights remain under cc-by-nc-4.0, with upstream terms in LICENSE. Non-commercial use only under the original CC BY-NC 4.0 terms. Converter/runtime additions are MIT licensed (CODE_LICENSE). No new task training, safety certification or commercial license is implied.

Other formats