CRZbulabula commented on code in PR #15586:
URL: https://github.com/apache/iotdb/pull/15586#discussion_r2110641448
##########
iotdb-core/ainode/ainode/core/manager/inference_manager.py:
##########
@@ -52,6 +52,17 @@ def infer(self, full_data, predict_length=96, **_):
df = pd.DataFrame(output[0])
return convert_to_binary(df)
+class SundialStrategy(InferenceStrategy):
+ def infer(self, full_data, predict_length=96, **_):
+ data = full_data[1][0]
+ if data.dtype.byteorder not in ('=', '|'):
+ data = data.byteswap().newbyteorder()
+ seqs = torch.tensor(data).unsqueeze(0).float()
Review Comment:
Yes, the current input are different, the demanding input class is tensor
rather than nparray cuz we invoke the `generate` func for sundial. Let's leave
the data processing logic here right now and unify this when we refactoring the
integration of TimerXL
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
