Add files using upload-large-folder tool

README.md

@@ -1,6 +1,5 @@
 ---
 license: cc-by-4.0
-library_name: onnx
 tags:
 - onnx
 - quantized
@@ -9,11 +8,6 @@ tags:
 - embeddings
 datasets:
 - nynefold/FreshRet-Small
-base_model:
-- perplexity-ai/pplx-embed-v1-0.6b
-- codefuse-ai/F2LLM-v2-1.7B
-- Qwen/Qwen3-Reranker-0.6B
-- zeroentropy/zerank-1-small
 ---
 # retrieval-quant: quantized ONNX retrievers
 
@@ -28,7 +22,8 @@ Repos: [`nynefold/pplx-embed-v1-0.6b-ONNX`](https://huggingface.co/nynefold/pplx
 [`nynefold/zerank-1-small-ONNX`](https://huggingface.co/nynefold/zerank-1-small-ONNX), and the evaluation set
 [`nynefold/FreshRet-Small`](https://huggingface.co/datasets/nynefold/FreshRet-Small). `pipeline_example.py` in this repo
 retrieves with pplx-embed and reranks with zerank, using `inference.py` (ONNX Runtime, `tokenizers`
-and NumPy only).
+and NumPy only). All six repos are in the collection
+[retrieval-quant: quantized ONNX retrievers](https://huggingface.co/collections/nynefold/retrieval-quant-quantized-onnx-retrievers-6adb9c3a7f6012982d819511).
 
 ## Results in short
 
@@ -223,7 +218,7 @@ Base model [`Qwen/Qwen3-Reranker-0.6B`](https://huggingface.co/Qwen/Qwen3-Rerank
 
 ### zerank-1-small
 
-Base model [`zeroentropy/zerank-1-small`](https://huggingface.co/zeroentropy/zerank-1-small), FP32 6885 MB. Files: [`nynefold/zerank-1-small-ONNX`](https://huggingface.co/nynefold/zerank-1-small-ONNX).
+Base model [`zeroentropy/zerank-1-small-reranker`](https://huggingface.co/zeroentropy/zerank-1-small-reranker), FP32 6885 MB. Files: [`nynefold/zerank-1-small-ONNX`](https://huggingface.co/nynefold/zerank-1-small-ONNX).
 
 | variant | MB | Spearman mean (gate 0.98) | FreshRet Small nDCG@10 vs FP32 (gate -1.0 pt) | verdict |
 |---|---:|---:|---|---|

inference.py

@@ -13,7 +13,7 @@ Needs only `onnxruntime`, `tokenizers` and `numpy` (plus `huggingface_hub` to do
     print(rr.rerank("how do I reset the router?", ["Hold the reset button ...", "The router has ..."]))
 
 `load` takes a repo id or a local folder. All model-specific details (prompts, truncation, output
-transform) come from `inference_config.json`, which is part of the contract: the graphs expect
+transform) come from `config.json`, which is part of the contract: the graphs expect
 exactly these inputs.
 """
 import json
@@ -24,7 +24,7 @@ import numpy as np
 import onnxruntime as ort
 from tokenizers import Tokenizer
 
-CONFIG = "inference_config.json"
+CONFIG = "config.json"  # also the file the Hub counts downloads by
 
 
 def _local(repo_or_dir, variant):