Some popular image models like Mask-RCNN for object detection calculate the loss in a slightly different way: loss_dict = model(images, targets) losses = sum(loss for loss in loss_dict.values()) other than out = model(images) loss = criteria(out, targets) Orca estimator wants to support this model....

https://github.com/intel-analytics/BigDL/actions/runs/4080727352/jobs/7035191801#step:6:7683

fsspec: https://github.com/intel-analytics/BigDL/blob/main/python/orca/src/bigdl/orca/learn/pytorch/torch_runner.py#L703 [already removed, not needed now: https://github.com//pull/7426]

https://github.com/intel-analytics/BigDL/blob/main/python/chronos/example/hpo/muti_objective_hpo_with_builtin_latency_tutorial.ipynb In above notebook when search space is changed from Categorical to Real optuna is throwing some error. Error Error persist even after updating optuna to latest versi...

Problem When I print the outputs of tf_estimator.predict(xshards) using backend="ray" and backend="spark", I find that they are inconsistent in the data structure. backend="ray" will output a list of x_y dict, while backend="spark" will output an XShards. The expected output data should be an XSha...

Problem When I want to save tensorflow model(which is a directory) to remote hdfs directory on yarn cluster(cluster_mode="yarn-client") using estimator.save(remote_model_path), an error occurs showing mkdir: permission denied. So I use TemporaryDirectory to save it to the local temporary directory a...

Env: i7 12th win11 python3.8.16 OV version: 2022.3.0-9052-9752fafe8eb-releases/2022/3 Error message:   File "c:\Users\yinachen\AppData\Local\miniconda3\envs\ldm\lib\site-packages\bigdl\nano\deps\openvino\core\utils.py", line 54, in convert_onnx_to_xml           p = subprocess.Popen(mo_cmd.split())  ...

According to the users' feedbacks and our own tests, we have encountered following issues when upgrade TensorFlow to 2.11.0 Deprecated API (get_weights), please refer to Issue #7319 . When running evaluation of listwise ranking example with multiple workers, the metrics accumulate as the number of...

When the input file is encrypted data, PPMLContext doesn't read and decrypt correctly. It outputs garbled code during SQL operation, which means that the ciphertext has not been decrypted correctly. Exception in thread "main" org.apache.spark.sql.AnalysisException: Cannot resolve column name "name"...

I would like to know if there are any plans to compile bigdl-dllib with Scala 2.13, and publish the artifact. Currently I´m working with Scala 3 in some research and I would like to present a paper for the next Scaladays using new lang features + some bigdl examples. Thanks!