crossmodalgroup / er-san Goto Github PK
View Code? Open in Web Editor NEWImplementation of our IJCAI2022 oral paper, ER-SAN: Enhanced-Adaptive Relation Self-Attention Network for Image Captioning.
Implementation of our IJCAI2022 oral paper, ER-SAN: Enhanced-Adaptive Relation Self-Attention Network for Image Captioning.
你好,我下载了你的训练日志,用tensorboard打开了,看见训练的CIDEr最高只有119.9,请问您的论文指标上添加了什么训练策略吗,谢谢,希望解答我的问题
image 275025: woods frames snout tulip caught dinosaurs woods peek
evaluating validation preformance... 4979/5000 (1.840761)
image 362971: woods golden ended union woods resemble china tablets
image 474: woods climber benched tulip musical draining bringing cooking launching
image 49327: woods hanging scraper woods sealed interviewed woods foreheads
image 144959: woods silver woody elegantly bringing crossroads honk bringing woods goofing
image 349414: woods hanging rice dinosaurs nicely china woods arrow bringing woods goofing
image 143359: woods hanging climate off cooking rags rich scraper woods sealed
image 546658: woods hanging rice dinosaurs cooking interstate service woods cafe
image 200563: woods hanging scraper woods cafe interviewed woods interstate sunroof successful
image 345469: woods guy pretending dinosaurs woods teddy sorted woody fisherman china knit
image 306619: woods majestically dinosaurs telescope gallops bundt woods buying door
evaluating validation preformance... 4989/5000 (1.860643)
image 324313: woods holds woody woods seating feels honk bringing retrieve china crossroads
image 46616: minnie memorial rice dozen stern woods peep bringing woods gain
image 285832: woods heavy starring woody woods leather parrots caddy want woods goofing
image 496718: woods climber benched musical woods overgrowth gains dinosaurs woods heavy starring
image 398209: woods heavy starring woody woods predators want woods goofing
image 568041: woods heavy starring woody woods telescope caddy want woods goofing woody horizontal
image 206596: woods climber singer catchers woods draining buddhist bringing woods peek
image 451949: minnie lipstick woody tightly pecking bog roadway woods mirrors
image 203138: woods hanging want woods patch benched rice dinosaurs nicely china woods fit
image 296759: minnie burrito storefronts queue born dinosaurs ribbons want address
evaluating validation preformance... -1/5000 (1.820787)
loading annotations into memory...
0:00:00.119141
creating index...
index created!
using 5000/5000 predictions
Loading and preparing results...
DONE (t=0.02s)
creating index...
index created!
tokenization...
E:\song\ER-SAN-main\pycocoevalcap\tokenizer E:\song\ER-SAN-main\pycocoevalcap\tokenizer\tmph3xikfjn
PTBTokenizer tokenized 307821 tokens at 2790538.92 tokens per second.
E:\song\ER-SAN-main\pycocoevalcap\tokenizer E:\song\ER-SAN-main\pycocoevalcap\tokenizer\tmpht6d5wp0
PTBTokenizer tokenized 53842 tokens at 850787.47 tokens per second.
setting up scorers...
computing Bleu score...
{'testlen': 48843, 'reflen': 48341, 'guess': [48843, 43843, 38843, 33843], 'correct': [106, 0, 0, 0]}
ratio: 1.010384559690511
Bleu_1: 0.002
Bleu_2: 0.000
Bleu_3: 0.000
Bleu_4: 0.000
computing METEOR score...
METEOR: 0.010
computing Rouge score...
ROUGE_L: 0.002
computing CIDEr score...
CIDEr: 0.001
computing SPICE score...
Parsing reference captions
Initiating Stanford parsing pipeline
[main] INFO edu.stanford.nlp.pipeline.StanfordCoreNLP - Adding annotator tokenize
[main] INFO edu.stanford.nlp.pipeline.TokenizerAnnotator - TokenizerAnnotator: No tokenizer type provided. Defaulting to PTBTokenizer.
[main] INFO edu.stanford.nlp.pipeline.StanfordCoreNLP - Adding annotator ssplit
[main] INFO edu.stanford.nlp.pipeline.StanfordCoreNLP - Adding annotator parse
[main] INFO edu.stanford.nlp.parser.common.ParserGrammar - Loading parser from serialized file edu/stanford/nlp/models/lexparser/englishPCFG.ser.gz ...
done [0.2 sec].
[main] INFO edu.stanford.nlp.pipeline.StanfordCoreNLP - Adding annotator lemma
[main] INFO edu.stanford.nlp.pipeline.StanfordCoreNLP - Adding annotator ner
Loading classifier from edu/stanford/nlp/models/ner/english.all.3class.distsim.crf.ser.gz ... done [0.6 sec].
Loading classifier from edu/stanford/nlp/models/ner/english.muc.7class.distsim.crf.ser.gz ... done [0.3 sec].
Loading classifier from edu/stanford/nlp/models/ner/english.conll.4class.distsim.crf.ser.gz ... done [0.3 sec].
Threads( StanfordCoreNLP ) [25.998 seconds]
Threads( StanfordCoreNLP ) [23.241 seconds]
Threads( StanfordCoreNLP ) [11.351 seconds]
Parsing test captions
Threads( StanfordCoreNLP ) [13.569 seconds]
SPICE evaluation took: 1.350 min
SPICE: 0.003
model saved to log_transformer_triplet\model.pth
Read data: 0.0279085636138916
iter 186000 (epoch 16), train_loss = 1.558, time/batch = 0.310
Read data: 0.030405282974243164
iter 186001 (epoch 16), train_loss = 1.423, time/batch = 0.086
Read data: 0.026419639587402344
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.