Small changes

README.md

@@ -53,7 +53,7 @@ result = aspire_bienc(**inputs)
 clsrep = result.last_hidden_state[:,0,:]
 ```
 
-**`aspire-biencoder-compsci-spec-full`**, can be used as follows: 1) Download the [`aspire-biencoder-biomed-scib-full.zip`](https://drive.google.com/file/d/1AHtzyEpyn7DeFYOdt86ik4n0tGaG5kMC/view?usp=sharing), and 2) Use it per this example usage script: [`aspire/examples/ex_aspire_bienc.py`](https://github.com/allenai/aspire/blob/main/examples/ex_aspire_bienc.py)
+**`aspire-biencoder-compsci-spec-full`**, can be used as follows: 1) Download the [`aspire-biencoder-compsci-spec-full.zip`](https://drive.google.com/file/d/1AHtzyEpyn7DeFYOdt86ik4n0tGaG5kMC/view?usp=sharing), and 2) Use it per this example usage script: [`aspire/examples/ex_aspire_bienc.py`](https://github.com/allenai/aspire/blob/main/examples/ex_aspire_bienc.py)
 
 ### Variable and metrics
 This model is evaluated on information retrieval datasets with document level queries. Performance here is reported on CSFCube (computer science/English). This is detailed on [github](https://github.com/allenai/aspire) and in our [paper](https://arxiv.org/abs/2111.08366). CSFCube presents a finer-grained query via selected sentences in a query abstract based on which a finer-grained retrieval must be made from candidate abstracts. The bi-encoder above ignores the finer grained query sentences and uses the whole abstract - this presents a baseline in the paper.
@@ -62,13 +62,13 @@ We rank documents by the L2 distance between the query and candidate documents.
 
 ### Evaluation results
 
-The released model `aspire-biencoder-compsci-spec` (and `aspire-biencoder-compsci-spec-full`) is compared against `allenai/specter`. `aspire-biencoder-compsci-spec`<sup>*</sup> is the performance reported in our paper by averaging over 3 re-runs of the model. The released models `aspire-biencoder-compsci-spec` and `aspire-biencoder-compsci-spec-full` are the single best run among the 3 re-runs.
+The released model `aspire-biencoder-compsci-spec` (and `aspire-biencoder-compsci-spec-full`) is compared against `allenai/specter`. `aspire-biencoder-compsci-spec-full`<sup>*</sup> is the performance reported in our paper by averaging over 3 re-runs of the model. The released models `aspire-biencoder-compsci-spec` and `aspire-biencoder-compsci-spec-full` are the single best run among the 3 re-runs.
 
 |                                             | CSFCube aggregated |    CSFCube aggregated|
 |--------------------------------------------:|:---------:|:-------:|
 |                                             |    MAP    | NDCG%20 |
 |                               `specter`     |   34.23   | 53.28   |
-| `aspire-biencoder-compsci-spec`<sup>*</sup> |  37.90    | 58.16   |
+| `aspire-biencoder-compsci-spec-full`<sup>*</sup> |  37.90    | 58.16   |
 |             `aspire-biencoder-compsci-spec` |  37.17    | 57.91   |
 |        `aspire-biencoder-compsci-spec-full` |  37.67    | 59.26   |