xinhaomei / dcase2021_task6_v2 Goto Github PK
View Code? Open in Web Editor NEWCode for CVSSP submission to DCASE 2021 Task 6
Code for CVSSP submission to DCASE 2021 Task 6
Hi, I am getting this error:
FileNotFoundError: [Errno 2] No such file or directory: 'data/Clotho/pickles/456/train_keywords_dict_pred_5.p'
what could cause this ? there is no train_keywords_dict_pred_x.p file. Is it missing or did I do something wrong ?
when i run the train.py , i encountered some difficulties,
I've tried re-downloading coco_caption to no avail, and I thought at first that there wasn't enough memory, and after re-running it I still got the same error.
Can you give me some advice?
Thank you
Traceback (most recent call last): File "train.py", line 323, in <module> eval_beam(evaluation_data, beam_size=2) File "train.py", line 142, in eval_beam beam_metrics = evaluate_metrics(captions_pred, captions_gt) File "/data/shaoxi/hty_all/task6_v2/eval_metrics.py", line 293, in evaluate_metrics metrics, per_file_metrics = evaluate_metrics_from_lists(predictions, ground_truths) File "/data/shaoxi/hty_all/task6_v2/eval_metrics.py", line 165, in evaluate_metrics_from_lists metrics, per_file_metrics = evaluate_metrics_from_files(pred_file, ref_file) File "/data/shaoxi/hty_all/task6_v2/eval_metrics.py", line 112, in evaluate_metrics_from_files cocoEval.evaluate(verbose=False) File "/data/shaoxi/hty_all/task6_v2/coco_caption/pycocoevalcap/eval.py", line 62, in evaluate score, scores = scorer.compute_score(gts, res) File "/data/shaoxi/hty_all/task6_v2/coco_caption/pycocoevalcap/spice/spice.py", line 75, in compute_score cwd=os.path.dirname(os.path.abspath(__file__))) File "/home/user1/.conda/envs/hty/lib/python3.7/subprocess.py", line 363, in check_call raise CalledProcessError(retcode, cmd) subprocess.CalledProcessError: Command '['java', '-jar', '-Xmx8G', 'spice-1.0.jar', '/data/shaoxi/hty_all/task6_v2/coco_caption/pycocoevalcap/spice/tmp/tmpzce3f7z7', '-cache', '/data/shaoxi/hty_all/task6_v2/coco_caption/pycocoevalcap/spice/cache', '-out', '/data/shaoxi/hty_all/task6_v2/coco_caption/pycocoevalcap/spice/tmp/tmpvpuknq04', '-subset', '-silent']' returned non-zero exit status 1.
Thanks again for the excellent work,
it is not clear to me how the settings.yaml
should be set to perform the first step you indicate in your work. How do you train your framework with Audiocaps?
Thanks in advance
Hi, thank you very much for your work!
I am reading your code with report to understand the work you did. How do you implement the reinforcement learning in "config.mode == 'finetune' "? Specifically, which part of the code is the 'CIDEr loss' part in 'Fine-tuning using RL'?
I saw that your parameter settings are very different from those of PANN.
The hop length = 512 , sample rate = 44100 in your project ; but in PANN ,these parameters are different from you.
Do these parameters affect the experimental results? and do I need to set the acoustic feature extraction parameters the same as Pann?
Thank you very much~
Hi,
In my opinion, model structure in this repo is the same with your ACT model, they both use a full Transformer network based on an encoder-decoder architecture.
This paper further uses transfer learning and reinforcement learning to fine tune the parameters, but why the result in this paper is worse than the ACT?
Take following two models as an example:
Model | BLEU1 | BLEU2 | BLEU3 | BLEU4 | ROUGEL | METERO | CIDEr | SPICE | SPIDEr |
---|---|---|---|---|---|---|---|---|---|
ACT_m_DeiT_AudioSet | 0.653 | 0.495 | 0.363 | 0.259 | 0.471 | 0.222 | 0.663 | 0.163 | 0.413 |
B+PANNs+AC+RL | 0.634 | 0.423 | 0.288 | 0.185 | 0.410 | 0.187 | 0.476 | 0.134 | 0.305 |
Is available the code for training the wav2vec model as described in the technical report or is just released in pre-trained models folder?
Thanks for sharing!
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.