prithivida
commited on
Commit
•
2f80623
1
Parent(s):
c1fd189
Update README.md
Browse files
README.md
CHANGED
@@ -45,7 +45,7 @@ pipeline_tag: sentence-similarity
|
|
45 |
|
46 |
- [License and Terms:](#license-and-terms)
|
47 |
- [Detailed comparison & Our Contribution:](#detailed-comparison--our-contribution)
|
48 |
-
- [ONNX & GGUF
|
49 |
- [Usage:](#usage)
|
50 |
- [With Sentence Transformers:](#with-sentence-transformers)
|
51 |
- [With Huggingface Transformers:](#with-huggingface-transformers)
|
@@ -91,6 +91,13 @@ Full set of evaluation numbers for our model
|
|
91 |
|
92 |
<br/>
|
93 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
94 |
# Usage:
|
95 |
|
96 |
#### With Sentence Transformers:
|
@@ -167,7 +174,7 @@ The below numbers are with mDPR model, but miniMiracle_zh_v1 should give a even
|
|
167 |
*Note: MIRACL paper shows a different (higher) value for BM25 Chinese, So we are taking that value from BGE-M3 paper, rest all are form the MIRACL paper.*
|
168 |
|
169 |
#### cMTEB numbers:
|
170 |
-
CMTEB is a general purpose embedding evaluation
|
171 |
We ran the retrieval slice of the cMTEB and add the scores here.
|
172 |
|
173 |
We compared the performance few top general purpose embedding models on the C-MTEB benchmark. please refer to the C-MTEB leaderboard.
|
|
|
45 |
|
46 |
- [License and Terms:](#license-and-terms)
|
47 |
- [Detailed comparison & Our Contribution:](#detailed-comparison--our-contribution)
|
48 |
+
- [ONNX & GGUF Status:](#onnx-gguf-status)
|
49 |
- [Usage:](#usage)
|
50 |
- [With Sentence Transformers:](#with-sentence-transformers)
|
51 |
- [With Huggingface Transformers:](#with-huggingface-transformers)
|
|
|
91 |
|
92 |
<br/>
|
93 |
|
94 |
+
# ONNX & GGUF Status:
|
95 |
+
|
96 |
+
|Variant| Status |
|
97 |
+
|:---:|:---:|
|
98 |
+
|FP16 ONNX | ✅ |
|
99 |
+
|GGUF | WIP|
|
100 |
+
|
101 |
# Usage:
|
102 |
|
103 |
#### With Sentence Transformers:
|
|
|
174 |
*Note: MIRACL paper shows a different (higher) value for BM25 Chinese, So we are taking that value from BGE-M3 paper, rest all are form the MIRACL paper.*
|
175 |
|
176 |
#### cMTEB numbers:
|
177 |
+
CMTEB is a general purpose embedding evaluation benchmark covering wide range of tasks, but like BGE-M3, miniMiracle models are predominantly tuned for retireval tasks aimed at search & IR based usecases.
|
178 |
We ran the retrieval slice of the cMTEB and add the scores here.
|
179 |
|
180 |
We compared the performance few top general purpose embedding models on the C-MTEB benchmark. please refer to the C-MTEB leaderboard.
|