Skip to content
TheChatArchiveThe Chat Archive
gpt-4-1106-preview· OpenAIAi2 WildChat-4.8M (AI2)

model = TimeSeriesKMeans(nclusters=nclusters, metric="dtw", verbose=True, njobs=-1)…

  1. User

    model = TimeSeriesKMeans(n_clusters=n_clusters, metric="dtw", verbose=True, n_jobs=-1) model.fit(time_series_data_scaled[i:i+batch_size])

    没有打印损失,输出的这是些什么东西,怎么这么多

    [Parallel(n_jobs=-1)]: Done 6000 out of 6000 | elapsed: 4.0s finished [Parallel(n_jobs=-1)]: Using backend ThreadingBackend with 16 concurrent workers. [Parallel(n_jobs=-1)]: Done 18 tasks | elapsed: 0.0s [Parallel(n_jobs=-1)]: Done 168 tasks | elapsed: 0.2s [Parallel(n_jobs=-1)]: Done 418 tasks | elapsed: 0.4s [Parallel(n_jobs=-1)]: Done 768 tasks | elapsed: 0.7s [Parallel(n_jobs=-1)]: Done 1218 tasks | elapsed: 1.0s [Parallel(n_jobs=-1)]: Done 1768 tasks | elapsed: 1.5s [Parallel(n_jobs=-1)]: Done 2418 tasks | elapsed: 2.0s [Parallel(n_jobs=-1)]: Done 3168 tasks | elapsed: 2.5s [Parallel(n_jobs=-1)]: Done 4018 tasks | elapsed: 3.2s [Parallel(n_jobs=-1)]: Done 4968 tasks | elapsed: 3.8s [Parallel(n_jobs=-1)]: Done 6000 out of 6000 | elapsed: 4.5s finished [Parallel(n_jobs=-1)]: Using backend ThreadingBackend with 16 concurrent workers. [Parallel(n_jobs=-1)]: Done 18 tasks | elapsed: 0.0s [Parallel(n_jobs=-1)]: Done 168 tasks | elapsed: 0.1s [Parallel(n_jobs=-1)]: Done 418 tasks | elapsed: 0.3s [Parallel(n_jobs=-1)]: Done 768 tasks | elapsed: 0.6s [Parallel(n_jobs=-1)]: Done 1218 tasks | elapsed: 0.9s [Parallel(n_jobs=-1)]: Done 1768 tasks | elapsed: 1.3s [Parallel(n_jobs=-1)]: Done 2418 tasks | elapsed: 1.8s [Parallel(n_jobs=-1)]: Done 3168 tasks | elapsed: 2.4s [Parallel(n_jobs=-1)]: Done 4018 tasks | elapsed: 3.0s [Parallel(n_jobs=-1)]: Done 4968 tasks | elapsed: 3.6s [Parallel(n_jobs=-1)]: Done 6000 out of 6000 | elapsed: 4.4s finished [Parallel(n_jobs=-1)]: Using backend ThreadingBackend with 16 concurrent workers. [Parallel(n_jobs=-1)]: Done 18 tasks | elapsed: 0.0s [Parallel(n_jobs=-1)]: Done 168 tasks | elapsed: 0.1s [Parallel(n_jobs=-1)]: Done 418 tasks | elapsed: 0.3s [Parallel(n_jobs=-1)]: Done 768 tasks | elapsed: 0.6s [Parallel(n_jobs=-1)]: Done 1218 tasks | elapsed: 1.0s [Parallel(n_jobs=-1)]: Done 1768 tasks | elapsed: 1.5s [Parallel(n_jobs=-1)]: Done 2418 tasks | elapsed: 2.2s [Parallel(n_jobs=-1)]: Done 3168 tasks | elapsed: 2.8s [Parallel(n_jobs=-1)]: Done 4018 tasks | elapsed: 3.3s [Parallel(n_jobs=-1)]: Done 4968 tasks | elapsed: 4.0s [Parallel(n_jobs=-1)]: Done 6000 out of 6000 | elapsed: 4.7s finished [Parallel(n_jobs=-1)]: Using backend ThreadingBackend with 16 concurrent workers. [Parallel(n_jobs=-1)]: Done 18 tasks | elapsed: 0.0s [Parallel(n_jobs=-1)]: Done 168 tasks | elapsed: 0.1s [Parallel(n_jobs=-1)]: Done 418 tasks | elapsed: 0.3s [Parallel(n_jobs=-1)]: Done 768 tasks | elapsed: 0.6s [Parallel(n_jobs=-1)]: Done 1218 tasks | elapsed: 1.0s [Parallel(n_jobs=-1)]: Done 1768 tasks | elapsed: 1.4s [Parallel(n_jobs=-1)]: Done 2418 tasks | elapsed: 1.9s [Parallel(n_jobs=-1)]: Done 3168 tasks | elapsed: 2.5s [Parallel(n_jobs=-1)]: Done 4018 tasks | elapsed: 3.1s [Parallel(n_jobs=-1)]: Done 4968 tasks | elapsed: 3.8s [Parallel(n_jobs=-1)]: Done 6000 out of 6000 | elapsed: 4.4s finished [Parallel(n_jobs=-1)]: Using backend ThreadingBackend with 16 concurrent workers. [Parallel(n_jobs=-1)]: Done 18 tasks | elapsed: 0.0s [Parallel(n_jobs=-1)]: Done 168 tasks | elapsed: 0.1s [Parallel(n_jobs=-1)]: Done 418 tasks | elapsed: 0.3s [Parallel(n_jobs=-1)]: Done 768 tasks | elapsed: 0.5s [Parallel(n_jobs=-1)]: Done 1218 tasks | elapsed: 0.8s [Parallel(n_jobs=-1)]: Done 1768 tasks | elapsed: 1.1s [Parallel(n_jobs=-1)]: Done 2418 tasks | elapsed: 1.5s [Parallel(n_jobs=-1)]: Done 3168 tasks | elapsed: 2.2s [Parallel(n_jobs=-1)]: Done 4018 tasks | elapsed: 2.8s [Parallel(n_jobs=-1)]: Done 4968 tasks | elapsed: 3.5s [Parallel(n_jobs=-1)]: Done 6000 out of 6000 | elapsed: 4.3s finished [Parallel(n_jobs=-1)]: Using backend ThreadingBackend with 16 concurrent workers. [Parallel(n_jobs=-1)]: Done 18 tasks | elapsed: 0.0s [Parallel(n_jobs=-1)]: Done 168 tasks | elapsed: 0.1s [Parallel(n_jobs=-1)]: Done 418 tasks | elapsed: 0.3s [Parallel(n_jobs=-1)]: Done 768 tasks | elapsed: 0.6s [Parallel(n_jobs=-1)]: Done 1218 tasks | elapsed: 0.9s [Parallel(n_jobs=-1)]: Done 1768 tasks | elapsed: 1.3s [Parallel(n_jobs=-1)]: Done 2418 tasks | elapsed: 1.7s [Parallel(n_jobs=-1)]: Done 3168 tasks | elapsed: 2.3s [Parallel(n_jobs=-1)]: Done 4018 tasks | elapsed: 3.0s [Parallel(n_jobs=-1)]: Done 4968 tasks | elapsed: 3.7s [Parallel(n_jobs=-1)]: Done 6000 out of 6000 | elapsed: 4.6s finished [Parallel(n_jobs=-1)]: Using backend ThreadingBackend with 16 concurrent workers. [Parallel(n_jobs=-1)]: Done 18 tasks | elapsed: 0.0s

  2. Assistant

    The output you're seeing is generated by the joblib library, which is used under the hood by TimeSeriesKMeans for parallel computation. TimeSeriesKMeans is a clustering algorithm from the tslearn package that is designed to group similar time series together. You are using Dynamic Time Warping (DTW) as the distance metric.

    When you pass verbose=True to the TimeSeriesKMeans constructor, you enable verbose mode, which means that various messages about the progress of the computation are printed during the fitting process.

    Here's what those output messages mean:

    • [Parallel(n_jobs=-1)]: This indicates that joblib is running in parallel mode, and n_jobs=-1 means that it's using all available CPU cores.

    • Done 6000 out of 6000 | elapsed: 4.0s finished: This message tells you that joblib has completed 6000 tasks (probably pairwise DTW computations or cluster assignments in this context), and that it took 4.0 seconds to finish all of them.

    • Using backend ThreadingBackend with 16 concurrent workers: This tells you that joblib is using a threading backend, which allows for parallelism through thread-level concurrency, and that it's using 16 threads to do the work.

    The repetition that you're seeing is likely because the algorithm has several steps where parallel computation is involved, and each of these steps is reporting its own progress. If you're fitting the model multiple times (like in a loop) or if the algorithm iterates over the data several times, these messages would appear repeatedly.

    If you do not want these verbose messages, you can set verbose=False. However, note that the verbosity here doesn't pertain directly to the loss of the TimeSeriesKMeans algorithm during its iterations, but rather to the progress of parallel computations.

    If you're interested in the loss or inertia (i.e., the sum of distances of samples to their closest cluster center), you might have to look at the attributes of the fitted model object after the fitting process, such as model.inertia_ if it's available in tslearn. Alternatively, you might also consider setting the verbose parameter to a different level if the API allows for more granularity in the messages displayed.