spacemanidol commited on
Commit
a8e7a46
1 Parent(s): 2040317

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +37 -0
README.md CHANGED
@@ -30,6 +30,43 @@ All embeddings are created using [Snowflake's Arctic Embed M v1.5](https://huggi
30
  It's worth noting that Snowflake's Arctic Embed M v1.5 is optimized for efficient embeddings and thus supports embedding truncation and quantization. More details on model release can be found in this [blog](https://www.snowflake.com/engineering-blog/arctic-embed-m-v1-5-enterprise-retrieval/) along with methods for [quantization and compression](https://github.com/Snowflake-Labs/arctic-embed/blob/main/compressed_embeddings_examples/score_arctic_embed_m_v1dot5_with_quantization.ipynb).
31
  Note, that the embeddings are not normalized so you will need to normalize them before usage.
32
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
33
  ## Loading the dataset
34
 
35
  ### Loading the document embeddings
 
30
  It's worth noting that Snowflake's Arctic Embed M v1.5 is optimized for efficient embeddings and thus supports embedding truncation and quantization. More details on model release can be found in this [blog](https://www.snowflake.com/engineering-blog/arctic-embed-m-v1-5-enterprise-retrieval/) along with methods for [quantization and compression](https://github.com/Snowflake-Labs/arctic-embed/blob/main/compressed_embeddings_examples/score_arctic_embed_m_v1dot5_with_quantization.ipynb).
31
  Note, that the embeddings are not normalized so you will need to normalize them before usage.
32
 
33
+
34
+ ## Retrieval Performance
35
+ Retrieval performance for the TREC DL21-23, MSMARCOV2-Dev and Raggy Queries can be found below with BM25 as a baseline. For both systems, retrieval is at the segment level and Doc Score = Max (passage score).
36
+ Retrieval is done via a dot product and happens in BF16. Since the M-v1.5 model supports Vector Truncation we do so to 256 dimensions
37
+
38
+ ### NDCG@10
39
+ | Dataset | BM25 | Arctic-M-V1.5 (768 Dimensions) | Arctic-M-V1.5 (256 Dimensions) |
40
+ |---|---|---|---|
41
+ | Deep Learning 2021 | 0.5778 | 0.6936 | 0.69392 |
42
+ | Deep Learning 2022 | 0.3576 | 0.55199 | 0.55608 |
43
+ | Deep Learning 2023 | 0.3356 | 0.46963 | 0.45196 |
44
+ | msmarcov2-dev | N/A | 0.346 | 0.34074 |
45
+ | msmarcov2-dev2 | N/A | 0.34518 | 0.34339 |
46
+ | Raggy Queries | 0.4227 | 0.57439 | 0.56686 |
47
+
48
+ ### Recall@100
49
+ | Dataset | BM25 | Arctic-M-V1.5 (768 Dimensions) | Arctic-M-V1.5 (256 Dimensions) |
50
+ |---|---|---|---|
51
+ | Deep Learning 2021 | 0.3811 | 0.43 | 0.42245 |
52
+ | Deep Learning 2022 | 0.233 | 0.32125 | 0.3165 |
53
+ | Deep Learning 2023 | 0.3049 | 0.37622 | 0.36089 |
54
+ | msmarcov2-dev | 0.6683 | 0.85435 | 0.84985 |
55
+ | msmarcov2-dev2 | 0.6771 | 0.8576 | 0.8526 |
56
+ | Raggy Queries | 0.2807 | 0.36915 | 0.36149 |
57
+
58
+
59
+ ### Recall@1000
60
+ | Dataset | BM25 | Arctic-M-V1.5 (768 Dimensions) | Arctic-M-V1.5 (256 Dimensions) |
61
+ |---|---|---|---|
62
+ | Deep Learning 2021 | 0.7115 | 0.74895 | 0.73511 |
63
+ | Deep Learning 2022 | 0.479 | 0.55413 | 0.54499 |
64
+ | Deep Learning 2023 | 0.5852 | 0.62262 | 0.61199 |
65
+ | msmarcov2-dev | 0.8528 | 0.94156 | 0.94014 |
66
+ | msmarcov2-dev2 | 0.8577 | 0.94277 | 0.94047 |
67
+ | Raggy Queries | 0.5745 | 0.64527 | 0.63826 |
68
+
69
+
70
  ## Loading the dataset
71
 
72
  ### Loading the document embeddings