Adding temperature scaling on Joiner logits: (#789)

* Adding temperature scaling on Joiner logits:

- T hard-coded to 2.0
- so far best result NCE 0.122 (still not so high)
    - the BPE scores were rescaled with 0.2 (but then also incorrect words
      get high confidence, visually reasonable histograms are for 0.5 scale)
    - BPE->WORD score merging done by min(.) function
      (tried also prob-product, and also arithmetic, geometric, harmonic mean)

- without temperature scaling (i.e. scale 1.0), the best NCE was 0.032 (here product merging was best)

Results seem consistent with: https://arxiv.org/abs/2110.15222

Everything tuned on a very-small set of 100 sentences with 813 words and 10.2% WER, a Czech model.

I also experimented with blank posteriors mixed into the BPE confidences,
but no NCE improvement found, so not pushing that.

Temperature scling added also to the Greedy search confidences.

* making `temperature_scale` configurable from outside

sherpa-onnx/python/sherpa_onnx/online_recognizer.py

@@ -58,6 +58,7 @@ class OnlineRecognizer(object):
         model_type: str = "",
         lm: str = "",
         lm_scale: float = 0.1,
+        temperature_scale: float = 2.0,
     ):
         """
         Please refer to
@@ -123,6 +124,10 @@ class OnlineRecognizer(object):
           hotwords_score:
             The hotword score of each token for biasing word/phrase. Used only if
             hotwords_file is given with modified_beam_search as decoding method.
+          temperature_scale:
+            Temperature scaling for output symbol confidence estiamation.
+            It affects only confidence values, the decoding uses the original
+            logits without temperature.
           provider:
             onnxruntime execution providers. Valid values are: cpu, cuda, coreml.
           model_type:
@@ -193,6 +198,7 @@ class OnlineRecognizer(object):
             hotwords_score=hotwords_score,
             hotwords_file=hotwords_file,
             blank_penalty=blank_penalty,
+            temperature_scale=temperature_scale,
         )
 
         self.recognizer = _Recognizer(recognizer_config)