Muennighoff
commited on
Commit
·
50e5fa7
1
Parent(s):
f9baa76
Add SGPT-125M-weightedmean-nli
Browse files- 1_Pooling/config.json +9 -0
- README.md +89 -0
- config.json +54 -0
- config_sentence_transformers.json +7 -0
- eval/similarity_evaluation_sts-dev_results.csv +12 -0
- merges.txt +0 -0
- modules.json +14 -0
- pytorch_model.bin +3 -0
- sentence_bert_config.json +4 -0
- similarity_evaluation_sts-test_results.csv +2 -0
- special_tokens_map.json +1 -0
- tokenizer.json +0 -0
- tokenizer_config.json +1 -0
- vocab.json +0 -0
1_Pooling/config.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"word_embedding_dimension": 768,
|
3 |
+
"pooling_mode_cls_token": false,
|
4 |
+
"pooling_mode_mean_tokens": false,
|
5 |
+
"pooling_mode_max_tokens": false,
|
6 |
+
"pooling_mode_mean_sqrt_len_tokens": false,
|
7 |
+
"pooling_mode_weightedmean_tokens": true,
|
8 |
+
"pooling_mode_lasttoken": false
|
9 |
+
}
|
README.md
ADDED
@@ -0,0 +1,89 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
pipeline_tag: sentence-similarity
|
3 |
+
tags:
|
4 |
+
- sentence-transformers
|
5 |
+
- feature-extraction
|
6 |
+
- sentence-similarity
|
7 |
+
---
|
8 |
+
|
9 |
+
# {MODEL_NAME}
|
10 |
+
|
11 |
+
This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search.
|
12 |
+
|
13 |
+
<!--- Describe your model here -->
|
14 |
+
|
15 |
+
## Usage (Sentence-Transformers)
|
16 |
+
|
17 |
+
Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:
|
18 |
+
|
19 |
+
```
|
20 |
+
pip install -U sentence-transformers
|
21 |
+
```
|
22 |
+
|
23 |
+
Then you can use the model like this:
|
24 |
+
|
25 |
+
```python
|
26 |
+
from sentence_transformers import SentenceTransformer
|
27 |
+
sentences = ["This is an example sentence", "Each sentence is converted"]
|
28 |
+
|
29 |
+
model = SentenceTransformer('{MODEL_NAME}')
|
30 |
+
embeddings = model.encode(sentences)
|
31 |
+
print(embeddings)
|
32 |
+
```
|
33 |
+
|
34 |
+
|
35 |
+
|
36 |
+
## Evaluation Results
|
37 |
+
|
38 |
+
<!--- Describe how your model was evaluated -->
|
39 |
+
|
40 |
+
For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name={MODEL_NAME})
|
41 |
+
|
42 |
+
|
43 |
+
## Training
|
44 |
+
The model was trained with the parameters:
|
45 |
+
|
46 |
+
**DataLoader**:
|
47 |
+
|
48 |
+
`sentence_transformers.datasets.NoDuplicatesDataLoader.NoDuplicatesDataLoader` of length 8807 with parameters:
|
49 |
+
```
|
50 |
+
{'batch_size': 64}
|
51 |
+
```
|
52 |
+
|
53 |
+
**Loss**:
|
54 |
+
|
55 |
+
`sentence_transformers.losses.MultipleNegativesRankingLoss.MultipleNegativesRankingLoss` with parameters:
|
56 |
+
```
|
57 |
+
{'scale': 20.0, 'similarity_fct': 'cos_sim'}
|
58 |
+
```
|
59 |
+
|
60 |
+
Parameters of the fit()-Method:
|
61 |
+
```
|
62 |
+
{
|
63 |
+
"epochs": 1,
|
64 |
+
"evaluation_steps": 880,
|
65 |
+
"evaluator": "sentence_transformers.evaluation.EmbeddingSimilarityEvaluator.EmbeddingSimilarityEvaluator",
|
66 |
+
"max_grad_norm": 1,
|
67 |
+
"optimizer_class": "<class 'transformers.optimization.AdamW'>",
|
68 |
+
"optimizer_params": {
|
69 |
+
"lr": 2e-05
|
70 |
+
},
|
71 |
+
"scheduler": "WarmupLinear",
|
72 |
+
"steps_per_epoch": null,
|
73 |
+
"warmup_steps": 881,
|
74 |
+
"weight_decay": 0.01
|
75 |
+
}
|
76 |
+
```
|
77 |
+
|
78 |
+
|
79 |
+
## Full Model Architecture
|
80 |
+
```
|
81 |
+
SentenceTransformer(
|
82 |
+
(0): Transformer({'max_seq_length': 75, 'do_lower_case': False}) with Transformer model: GPTNeoModel
|
83 |
+
(1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': True, 'pooling_mode_lasttoken': False})
|
84 |
+
)
|
85 |
+
```
|
86 |
+
|
87 |
+
## Citing & Authors
|
88 |
+
|
89 |
+
<!--- Describe where people can find more information -->
|
config.json
ADDED
@@ -0,0 +1,54 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "EleutherAI/gpt-neo-125M",
|
3 |
+
"activation_function": "gelu_new",
|
4 |
+
"architectures": [
|
5 |
+
"GPTNeoModel"
|
6 |
+
],
|
7 |
+
"attention_dropout": 0,
|
8 |
+
"attention_layers": [
|
9 |
+
"global",
|
10 |
+
"local",
|
11 |
+
"global",
|
12 |
+
"local",
|
13 |
+
"global",
|
14 |
+
"local",
|
15 |
+
"global",
|
16 |
+
"local",
|
17 |
+
"global",
|
18 |
+
"local",
|
19 |
+
"global",
|
20 |
+
"local"
|
21 |
+
],
|
22 |
+
"attention_types": [
|
23 |
+
[
|
24 |
+
[
|
25 |
+
"global",
|
26 |
+
"local"
|
27 |
+
],
|
28 |
+
6
|
29 |
+
]
|
30 |
+
],
|
31 |
+
"bos_token_id": 50256,
|
32 |
+
"embed_dropout": 0,
|
33 |
+
"eos_token_id": 50256,
|
34 |
+
"gradient_checkpointing": false,
|
35 |
+
"hidden_size": 768,
|
36 |
+
"initializer_range": 0.02,
|
37 |
+
"intermediate_size": null,
|
38 |
+
"layer_norm_epsilon": 1e-05,
|
39 |
+
"max_position_embeddings": 2048,
|
40 |
+
"model_type": "gpt_neo",
|
41 |
+
"num_heads": 12,
|
42 |
+
"num_layers": 12,
|
43 |
+
"resid_dropout": 0,
|
44 |
+
"summary_activation": null,
|
45 |
+
"summary_first_dropout": 0.1,
|
46 |
+
"summary_proj_to_labels": true,
|
47 |
+
"summary_type": "cls_index",
|
48 |
+
"summary_use_proj": true,
|
49 |
+
"torch_dtype": "float32",
|
50 |
+
"transformers_version": "4.12.3",
|
51 |
+
"use_cache": true,
|
52 |
+
"vocab_size": 50257,
|
53 |
+
"window_size": 256
|
54 |
+
}
|
config_sentence_transformers.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"__version__": {
|
3 |
+
"sentence_transformers": "2.1.0",
|
4 |
+
"transformers": "4.12.3",
|
5 |
+
"pytorch": "1.10.0+cu113"
|
6 |
+
}
|
7 |
+
}
|
eval/similarity_evaluation_sts-dev_results.csv
ADDED
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
epoch,steps,cosine_pearson,cosine_spearman,euclidean_pearson,euclidean_spearman,manhattan_pearson,manhattan_spearman,dot_pearson,dot_spearman
|
2 |
+
0,880,0.8020888043867218,0.8061576061600323,0.8115790193917091,0.8127566445219281,0.8151774828346224,0.8168768253012519,0.6714352439215443,0.6842698031827484
|
3 |
+
0,1760,0.8207890933729857,0.8299220585031948,0.8229115620268723,0.8260390487060506,0.8226431671221693,0.8268412859144874,0.7163687537981843,0.7241325920123143
|
4 |
+
0,2640,0.825807966496274,0.8350680389228182,0.8234698522919874,0.82669509364278,0.8226906072899116,0.8273350157925823,0.7174974042697031,0.7223748348175156
|
5 |
+
0,3520,0.8239008360148115,0.8340749866360788,0.8204378001303088,0.8247398360656778,0.8194688522274942,0.8249822846315708,0.7074135254307587,0.7189552025643526
|
6 |
+
0,4400,0.828566413065448,0.838234779397801,0.8180607958439993,0.8235173992790007,0.8151299335266653,0.8219389594185353,0.7183458006365494,0.7267032968793666
|
7 |
+
0,5280,0.827628787266679,0.836641233958675,0.8155687572582913,0.8204596124712804,0.812573944370099,0.8184726342317181,0.7169432581111657,0.7275107589373734
|
8 |
+
0,6160,0.8256562468731141,0.8364368689027766,0.8142044319540191,0.8202859504297262,0.810812912007362,0.8181542941543771,0.7138174128230643,0.723863232641828
|
9 |
+
0,7040,0.827907976617676,0.8380915570236152,0.8139984331787357,0.8197369340366092,0.8103420760829132,0.817100964245968,0.7173747389896327,0.7271894537984986
|
10 |
+
0,7920,0.8304938445461576,0.840627480481868,0.8159166434702205,0.8219052931842625,0.8122654713176684,0.8191305457177783,0.7194307596248584,0.7307401057150337
|
11 |
+
0,8800,0.8298858387918985,0.8396322324640759,0.8137695661012616,0.819511857368826,0.8100806090497132,0.8170160478172596,0.7172998584064018,0.729081606349845
|
12 |
+
0,-1,0.8298576315342141,0.8396312822854843,0.8137206972008266,0.8194350868532124,0.8100226832521555,0.8169503876295982,0.7172951577499687,0.7291001882132022
|
merges.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
modules.json
ADDED
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
[
|
2 |
+
{
|
3 |
+
"idx": 0,
|
4 |
+
"name": "0",
|
5 |
+
"path": "",
|
6 |
+
"type": "sentence_transformers.models.Transformer"
|
7 |
+
},
|
8 |
+
{
|
9 |
+
"idx": 1,
|
10 |
+
"name": "1",
|
11 |
+
"path": "1_Pooling",
|
12 |
+
"type": "pooling.Pooling"
|
13 |
+
}
|
14 |
+
]
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:02c858688c428ccaae7f59a8d99b35939c4d8cb6956f1202324a22d1f220e579
|
3 |
+
size 551190545
|
sentence_bert_config.json
ADDED
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"max_seq_length": 75,
|
3 |
+
"do_lower_case": false
|
4 |
+
}
|
similarity_evaluation_sts-test_results.csv
ADDED
@@ -0,0 +1,2 @@
|
|
|
|
|
|
|
1 |
+
epoch,steps,cosine_pearson,cosine_spearman,euclidean_pearson,euclidean_spearman,manhattan_pearson,manhattan_spearman,dot_pearson,dot_spearman
|
2 |
+
-1,-1,0.7963431517583529,0.8095957063712698,0.7825362192798908,0.7848340068140034,0.779188870645588,0.7832460141326807,0.6250259237864847,0.6121175724719421
|
special_tokens_map.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"bos_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, "eos_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, "unk_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, "pad_token": "<|endoftext|>"}
|
tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer_config.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"unk_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "bos_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "eos_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "add_prefix_space": false, "errors": "replace", "model_max_length": 2048, "special_tokens_map_file": null, "name_or_path": "EleutherAI/gpt-neo-125M", "tokenizer_class": "GPT2Tokenizer"}
|
vocab.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|