mindmeld
mindmeld copied to clipboard
cisco
An Open Source Conversational AI Platform for Deep-Domain Voice Interfaces and Chatbots.
The default text preparation pipeline does not do any [unicode character normalization](https://docs.python.org/3/library/unicodedata.html#unicodedata.normalize), while a few of the default normalizer regexes do not include the combining diacritics in the alphabetic/word-like ranges....
The ElasticSearch versions that MM runs tests on [are no longer supported](https://www.elastic.co/support/eol). We need to move away from this and support more later versions.
Say an intent can only happen after another one in a domain? (cannot be the entering ones. ) another scenario: if we want to greet in two domain? how should...
This updates our Elasticsearch dependency in the dependency containers to mitigate https://www.cve.org/CVERecord?id=CVE-2021-44228.
So as part of attempting to solve #422 (but not doing so because of [the issue with MindMeld downloading an x86 build of Duckling from a bucket](https://github.com/cisco/mindmeld/issues/422#issuecomment-1160668153), I refactored the...
# Issue I decided that I wanted to try to run MindMeld on a Raspberry Pi and quickly found out that attempting to run `pip install mindmeld` in an `arm64`...
For active learning, there a couple of strategies that we currently utilize using the CRF suite model's marginal probabilities. [Studies](https://dl.acm.org/doi/pdf/10.5555/1613715.1613855) have shown that returning sequence-level probabilities instead of token-level marginal...
Currently most tagger models have initialization of the models happening in the set_params() method. This ideally needs to move into an __init__ function or to the setup_model() depending on what...
Default encoding differs across platforms (i.e. cp1252 in windows) which causes unexpected behaviours with certain character sets. It is suggested to specify the encoding type directly.
