tomaarsen HF Staff commited on
Commit
d166fa8
·
verified ·
1 Parent(s): 5ee8c7f

Fix CPU benchmark (ran on fp32 this time)

Browse files
Files changed (1) hide show
  1. README.md +18 -18
README.md CHANGED
@@ -449,24 +449,24 @@ All six released models were benchmarked against thirteen public rerankers on th
449
 
450
  | Model | Params | Best attn | pairs / second |
451
  |---|---:|---|---:|
452
- | **[`cross-encoder/ettin-reranker-17m-v1`](https://huggingface.co/cross-encoder/ettin-reranker-17m-v1)** | **17M** | SDPA | **76.1** |
453
- | [`cross-encoder/ms-marco-MiniLM-L4-v2`](https://huggingface.co/cross-encoder/ms-marco-MiniLM-L4-v2) | 19M | SDPA | 53.0 |
454
- | [`cross-encoder/ms-marco-MiniLM-L6-v2`](https://huggingface.co/cross-encoder/ms-marco-MiniLM-L6-v2) | 22M | SDPA | 29.4 |
455
- | **[`cross-encoder/ettin-reranker-32m-v1`](https://huggingface.co/cross-encoder/ettin-reranker-32m-v1)** | **32M** | SDPA | **28.5** |
456
- | [`cross-encoder/ms-marco-MiniLM-L12-v2`](https://huggingface.co/cross-encoder/ms-marco-MiniLM-L12-v2) | 33M | SDPA | 17.3 |
457
- | **[`cross-encoder/ettin-reranker-68m-v1`](https://huggingface.co/cross-encoder/ettin-reranker-68m-v1)** | **68M** | SDPA | **8.5** |
458
- | [`mixedbread-ai/mxbai-rerank-xsmall-v1`](https://huggingface.co/mixedbread-ai/mxbai-rerank-xsmall-v1) | 70M | eager | 6.0 |
459
- | [`BAAI/bge-reranker-base`](https://huggingface.co/BAAI/bge-reranker-base) | 278M | SDPA | 4.7 |
460
- | [`Alibaba-NLP/gte-reranker-modernbert-base`](https://huggingface.co/Alibaba-NLP/gte-reranker-modernbert-base) | 150M | SDPA | 3.7 |
461
- | **[`cross-encoder/ettin-reranker-150m-v1`](https://huggingface.co/cross-encoder/ettin-reranker-150m-v1)** | **150M** | SDPA | **3.6** |
462
- | [`ibm-granite/granite-embedding-reranker-english-r2`](https://huggingface.co/ibm-granite/granite-embedding-reranker-english-r2) | 150M | SDPA | 3.6 |
463
- | [`mixedbread-ai/mxbai-rerank-base-v1`](https://huggingface.co/mixedbread-ai/mxbai-rerank-base-v1) | 184M | eager | 2.4 |
464
- | **[`cross-encoder/ettin-reranker-400m-v1`](https://huggingface.co/cross-encoder/ettin-reranker-400m-v1)** | **400M** | SDPA | **1.3** |
465
- | [`BAAI/bge-reranker-large`](https://huggingface.co/BAAI/bge-reranker-large) | 560M | SDPA | 1.2 |
466
- | [`BAAI/bge-reranker-v2-m3`](https://huggingface.co/BAAI/bge-reranker-v2-m3) | 568M | SDPA | 1.2 |
467
- | [`mixedbread-ai/mxbai-rerank-base-v2`](https://huggingface.co/mixedbread-ai/mxbai-rerank-base-v2) | 494M | SDPA | 0.8 |
468
- | [`mixedbread-ai/mxbai-rerank-large-v1`](https://huggingface.co/mixedbread-ai/mxbai-rerank-large-v1) | 435M | eager | 0.8 |
469
- | **[`cross-encoder/ettin-reranker-1b-v1`](https://huggingface.co/cross-encoder/ettin-reranker-1b-v1)** | **1B** | SDPA | **0.5** |
470
 
471
  </details>
472
 
 
449
 
450
  | Model | Params | Best attn | pairs / second |
451
  |---|---:|---|---:|
452
+ | **[`cross-encoder/ettin-reranker-17m-v1`](https://huggingface.co/cross-encoder/ettin-reranker-17m-v1)** | **17M** | SDPA | **267.4** |
453
+ | [`cross-encoder/ms-marco-MiniLM-L4-v2`](https://huggingface.co/cross-encoder/ms-marco-MiniLM-L4-v2) | 19M | SDPA | 206.2 |
454
+ | [`cross-encoder/ms-marco-MiniLM-L6-v2`](https://huggingface.co/cross-encoder/ms-marco-MiniLM-L6-v2) | 22M | SDPA | 143.9 |
455
+ | **[`cross-encoder/ettin-reranker-32m-v1`](https://huggingface.co/cross-encoder/ettin-reranker-32m-v1)** | **32M** | SDPA | **92.5** |
456
+ | [`cross-encoder/ms-marco-MiniLM-L12-v2`](https://huggingface.co/cross-encoder/ms-marco-MiniLM-L12-v2) | 33M | SDPA | 75.9 |
457
+ | [`mixedbread-ai/mxbai-rerank-xsmall-v1`](https://huggingface.co/mixedbread-ai/mxbai-rerank-xsmall-v1) | 70M | eager | 38.9 |
458
+ | **[`cross-encoder/ettin-reranker-68m-v1`](https://huggingface.co/cross-encoder/ettin-reranker-68m-v1)** | **68M** | SDPA | **31.2** |
459
+ | [`BAAI/bge-reranker-base`](https://huggingface.co/BAAI/bge-reranker-base) | 278M | SDPA | 19.2 |
460
+ | [`Alibaba-NLP/gte-reranker-modernbert-base`](https://huggingface.co/Alibaba-NLP/gte-reranker-modernbert-base) | 150M | SDPA | 14.7 |
461
+ | [`ibm-granite/granite-embedding-reranker-english-r2`](https://huggingface.co/ibm-granite/granite-embedding-reranker-english-r2) | 150M | SDPA | 14.5 |
462
+ | **[`cross-encoder/ettin-reranker-150m-v1`](https://huggingface.co/cross-encoder/ettin-reranker-150m-v1)** | **150M** | SDPA | **14.0** |
463
+ | [`mixedbread-ai/mxbai-rerank-base-v1`](https://huggingface.co/mixedbread-ai/mxbai-rerank-base-v1) | 184M | eager | 13.4 |
464
+ | [`BAAI/bge-reranker-large`](https://huggingface.co/BAAI/bge-reranker-large) | 560M | SDPA | 6.2 |
465
+ | [`BAAI/bge-reranker-v2-m3`](https://huggingface.co/BAAI/bge-reranker-v2-m3) | 568M | SDPA | 6.0 |
466
+ | **[`cross-encoder/ettin-reranker-400m-v1`](https://huggingface.co/cross-encoder/ettin-reranker-400m-v1)** | **400M** | SDPA | **5.2** |
467
+ | [`mixedbread-ai/mxbai-rerank-large-v1`](https://huggingface.co/mixedbread-ai/mxbai-rerank-large-v1) | 435M | eager | 4.3 |
468
+ | [`mixedbread-ai/mxbai-rerank-base-v2`](https://huggingface.co/mixedbread-ai/mxbai-rerank-base-v2) | 494M | SDPA | 3.5 |
469
+ | **[`cross-encoder/ettin-reranker-1b-v1`](https://huggingface.co/cross-encoder/ettin-reranker-1b-v1)** | **1B** | SDPA | **2.1** |
470
 
471
  </details>
472