Muennighoff
commited on
Commit
•
7ce2897
1
Parent(s):
3eabd75
Better model with bs=1024
Browse files- README.md +1 -1
- config.json +1 -1
- config_sentence_transformers.json +2 -2
- eval/SGPT-1.3B-weightedmean-nli-bitfit_weightedmean_layer-1_results_average_precision.json +7 -0
- eval/SGPT-1.3B-weightedmean-nli-bitfit_weightedmean_layer-1_results_detailed.json +66 -0
- eval/quora.json +1 -0
- eval/similarity_evaluation_sts-dev_results.csv +11 -11
- pytorch_model.bin +2 -2
- tokenizer.json +0 -0
- tokenizer_config.json +1 -1
README.md
CHANGED
@@ -14,7 +14,7 @@ For usage instructions, refer to our codebase: https://github.com/Muennighoff/sg
|
|
14 |
|
15 |
## Evaluation Results
|
16 |
|
17 |
-
For eval results, refer to our paper: https://arxiv.org/abs/2202.08904
|
18 |
|
19 |
## Training
|
20 |
The model was trained with the parameters:
|
|
|
14 |
|
15 |
## Evaluation Results
|
16 |
|
17 |
+
For eval results, refer to the eval folder or our paper: https://arxiv.org/abs/2202.08904
|
18 |
|
19 |
## Training
|
20 |
The model was trained with the parameters:
|
config.json
CHANGED
@@ -67,7 +67,7 @@
|
|
67 |
},
|
68 |
"tokenizer_class": "GPT2Tokenizer",
|
69 |
"torch_dtype": "float32",
|
70 |
-
"transformers_version": "4.
|
71 |
"use_cache": true,
|
72 |
"vocab_size": 50257,
|
73 |
"window_size": 256
|
|
|
67 |
},
|
68 |
"tokenizer_class": "GPT2Tokenizer",
|
69 |
"torch_dtype": "float32",
|
70 |
+
"transformers_version": "4.20.0.dev0",
|
71 |
"use_cache": true,
|
72 |
"vocab_size": 50257,
|
73 |
"window_size": 256
|
config_sentence_transformers.json
CHANGED
@@ -1,7 +1,7 @@
|
|
1 |
{
|
2 |
"__version__": {
|
3 |
"sentence_transformers": "2.1.0",
|
4 |
-
"transformers": "4.
|
5 |
-
"pytorch": "1.10.
|
6 |
}
|
7 |
}
|
|
|
1 |
{
|
2 |
"__version__": {
|
3 |
"sentence_transformers": "2.1.0",
|
4 |
+
"transformers": "4.20.0.dev0",
|
5 |
+
"pytorch": "1.10.2"
|
6 |
}
|
7 |
}
|
eval/SGPT-1.3B-weightedmean-nli-bitfit_weightedmean_layer-1_results_average_precision.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"askubuntu": 55.97,
|
3 |
+
"cqadupstack": 13.47,
|
4 |
+
"twitterpara": 73.06,
|
5 |
+
"scidocs": 72.77,
|
6 |
+
"avg": 53.817499999999995
|
7 |
+
}
|
eval/SGPT-1.3B-weightedmean-nli-bitfit_weightedmean_layer-1_results_detailed.json
ADDED
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"askubuntu": {
|
3 |
+
"map_askubuntu_title": 55.97,
|
4 |
+
"p@1_askubuntu_title": 52.69,
|
5 |
+
"p@5_askubuntu_title": 41.94,
|
6 |
+
"mrr_askubuntu_title": 68.08
|
7 |
+
},
|
8 |
+
"cqadupstack": {
|
9 |
+
"map@100_cqadupstack_unix": 11.88,
|
10 |
+
"ndcg@10_cqadupstack_unix": 13.74,
|
11 |
+
"map@100_cqadupstack_gaming": 29.28,
|
12 |
+
"ndcg@10_cqadupstack_gaming": 31.75,
|
13 |
+
"map@100_cqadupstack_wordpress": 4.34,
|
14 |
+
"ndcg@10_cqadupstack_wordpress": 5.12,
|
15 |
+
"map@100_cqadupstack_stats": 14.68,
|
16 |
+
"ndcg@10_cqadupstack_stats": 16.12,
|
17 |
+
"map@100_cqadupstack_tex": 8.04,
|
18 |
+
"ndcg@10_cqadupstack_tex": 8.43,
|
19 |
+
"map@100_cqadupstack_english": 14.07,
|
20 |
+
"ndcg@10_cqadupstack_english": 15.59,
|
21 |
+
"map@100_cqadupstack_programmers": 10.78,
|
22 |
+
"ndcg@10_cqadupstack_programmers": 11.23,
|
23 |
+
"map@100_cqadupstack_mathematica": 10.74,
|
24 |
+
"ndcg@10_cqadupstack_mathematica": 12.57,
|
25 |
+
"map@100_cqadupstack_physics": 16.04,
|
26 |
+
"ndcg@10_cqadupstack_physics": 17.58,
|
27 |
+
"map@100_cqadupstack_gis": 14.81,
|
28 |
+
"ndcg@10_cqadupstack_gis": 16.19,
|
29 |
+
"map@100_cqadupstack_webmasters": 9.84,
|
30 |
+
"ndcg@10_cqadupstack_webmasters": 10.36,
|
31 |
+
"map@100_cqadupstack_android": 17.19,
|
32 |
+
"ndcg@10_cqadupstack_android": 19.08,
|
33 |
+
"map@100_cqadupstack_avg": 13.47,
|
34 |
+
"ndcg@10_cqadupstack_avg": 14.81
|
35 |
+
},
|
36 |
+
"twitterpara": {
|
37 |
+
"ap_twitter_twitterurl": 75.43,
|
38 |
+
"spearman_twitter_twitterurl": 70.6,
|
39 |
+
"ap_twitter_pit": 70.69,
|
40 |
+
"spearman_twitter_pit": 55.71,
|
41 |
+
"ap_twitter_avg": 73.06,
|
42 |
+
"spearman_twitter_avg": 63.15
|
43 |
+
},
|
44 |
+
"scidocs": {
|
45 |
+
"map_scidocs_cite_euclidean": 70.1,
|
46 |
+
"ndcg_scidocs_cite_euclidean": 85.17,
|
47 |
+
"map_scidocs_cite_cosine": 70.1,
|
48 |
+
"ndcg_scidocs_cite_cosine": 85.17,
|
49 |
+
"map_scidocs_cocite_euclidean": 72.87,
|
50 |
+
"ndcg_scidocs_cocite_euclidean": 86.72,
|
51 |
+
"map_scidocs_cocite_cosine": 72.87,
|
52 |
+
"ndcg_scidocs_cocite_cosine": 86.72,
|
53 |
+
"map_scidocs_coview_euclidean": 74.95,
|
54 |
+
"ndcg_scidocs_coview_euclidean": 87.03,
|
55 |
+
"map_scidocs_coview_cosine": 74.95,
|
56 |
+
"ndcg_scidocs_coview_cosine": 87.03,
|
57 |
+
"map_scidocs_coread_euclidean": 73.15,
|
58 |
+
"ndcg_scidocs_coread_euclidean": 86.15,
|
59 |
+
"map_scidocs_coread_cosine": 73.15,
|
60 |
+
"ndcg_scidocs_coread_cosine": 86.15,
|
61 |
+
"map_scidocs_euclidean_avg": 72.77,
|
62 |
+
"ndcg_scidocs_euclidean_avg": 86.27,
|
63 |
+
"map_scidocs_cosine_avg": 72.77,
|
64 |
+
"ndcg_scidocs_cosine_avg": 86.27
|
65 |
+
}
|
66 |
+
}
|
eval/quora.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"SGPT-1.3B-weightedmean-nli-bitfit": {"quora": {"NDCG@1": 0.7423, "NDCG@3": 0.78936, "NDCG@5": 0.80689, "NDCG@10": 0.8233, "NDCG@100": 0.84217, "NDCG@1000": 0.84504}}}
|
eval/similarity_evaluation_sts-dev_results.csv
CHANGED
@@ -1,12 +1,12 @@
|
|
1 |
epoch,steps,cosine_pearson,cosine_spearman,euclidean_pearson,euclidean_spearman,manhattan_pearson,manhattan_spearman,dot_pearson,dot_spearman
|
2 |
-
0,
|
3 |
-
0,
|
4 |
-
0,
|
5 |
-
0,
|
6 |
-
0,
|
7 |
-
0,
|
8 |
-
0,
|
9 |
-
0,
|
10 |
-
0,
|
11 |
-
0,
|
12 |
-
0,-1,0.
|
|
|
1 |
epoch,steps,cosine_pearson,cosine_spearman,euclidean_pearson,euclidean_spearman,manhattan_pearson,manhattan_spearman,dot_pearson,dot_spearman
|
2 |
+
0,440,0.8486729726641167,0.8514738189278961,0.8515245319252214,0.8509027260070884,0.8540666757323956,0.8537290162137693,0.7630773568229123,0.7600358216618573
|
3 |
+
0,880,0.860418177580361,0.8656263802961531,0.859274657391518,0.8610670106016408,0.8614574446998846,0.8635698806563004,0.775971737698288,0.7738330802938131
|
4 |
+
0,1320,0.8646586808912774,0.8708047895637789,0.8600953037639533,0.8630166525165047,0.8623501857502746,0.8655507432332792,0.774750540444686,0.7725645527279972
|
5 |
+
0,1760,0.8628892948335536,0.8689022188557769,0.8611001993825963,0.8627626879284295,0.8634677606209161,0.8653777840391851,0.7684411984607661,0.7660388115697093
|
6 |
+
0,2200,0.8623264709023419,0.8684038857716583,0.8599140937391133,0.8623163529776595,0.8621011652259446,0.8648639001116789,0.7658429961100489,0.7630158219390278
|
7 |
+
0,2640,0.8633934631804789,0.8700248299507874,0.8583066084846345,0.8610312177479946,0.8604516944303623,0.8632651971720137,0.7708747899267866,0.7695161449303083
|
8 |
+
0,3080,0.864875512993908,0.8708072415227665,0.8570051556310841,0.8598815222390387,0.8592210627649711,0.8622960844035745,0.7675942152106912,0.765566539001796
|
9 |
+
0,3520,0.8664358867199037,0.8717878560785026,0.8577414799031283,0.8608674339514554,0.8600366294063135,0.8633069107239323,0.7678182184536243,0.7655516315181986
|
10 |
+
0,3960,0.8660615852729263,0.8715945872618516,0.8580011946328364,0.8608961297007961,0.8603219261222281,0.8635694785207915,0.7658642808961628,0.7634980873963996
|
11 |
+
0,4400,0.8660032674381255,0.8715157046451364,0.8576564771305891,0.8606430352200829,0.8599938864592154,0.8633518022139872,0.7663838558727445,0.7645750276413869
|
12 |
+
0,-1,0.865995009654422,0.8715109608696208,0.857644450885013,0.8606063092160902,0.8599858692389015,0.8633254320890273,0.7663788803033962,0.7645777465044731
|
pytorch_model.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9e05f11a38a0abb8e97717b109e2d346a74cf1244e419ac5659416d9874487c8
|
3 |
+
size 5363081601
|
tokenizer.json
CHANGED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer_config.json
CHANGED
@@ -1 +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, "model_max_length": 2048, "special_tokens_map_file": null, "name_or_path": "EleutherAI/gpt-neo-1.3B", "errors": "replace", "tokenizer_class": "GPT2Tokenizer"}
|
|
|
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, "model_max_length": 2048, "special_tokens_map_file": null, "name_or_path": "EleutherAI/gpt-neo-1.3B", "errors": "replace", "pad_token": null, "add_bos_token": false, "tokenizer_class": "GPT2Tokenizer"}
|