Making the calls to the corresponding model from Flask was actually easier on AWS, since they were loaded into memory. Unfortunately, the scaling issues/excessive load times were big enough of an issue that we had to…
It probably would - however we built this out with a team of 1-2 engineers, so paying a bit more and dealing with maybe 20 seconds of extra latency a day is a price we are willing to pay for expedited/easier deployment.…
A bit of both. Flask was obviously not designed with serving Tensorflow models locally in mind, which is how we had it set up in v0. Towards the beginning we had to debug some weird threading issues, but towards the end…
Not inherently, but how we're using it makes it cheaper. Our stack has ~50 different ML models being served live, and GCP makes it easy to treat each model as a micro-service, and give auto-scaling to each one. This is…
While AWS does allow more hands-on control of the model stack, our "Custom Models" are more just different sets of weights using the exact same methodology. Basically, each customer that creates their own custom model…
For fine tuning BERT onto a specific domain, what amount of text data would you recommend to train on?
That depends on what you mean by noise. In the case of text classification, a lot of the noise in training is disagreement between human labelings. Unfortunately, classification systems will only be as good as the…
That is a good point. It helps to think of Precision and Recall (so by extension F1-Score) from your test data as random variables sampled from a distribution modeling the probability of getting each value in your…
I would imagine that AWS Comprehend utilizes a lot of customer feedback when training its internal models. However, they do advertise Comprehend as a general purpose NLP service, so they may have taken steps to make the…
Transfer learning is the premise of learning from data in one domain, and then being able to transfer that knowledge to another domain. In this case, we leverage large existing text datasets to build an initial model,…
There is definitely a feedback loop at play in the NLP community. Since most NLP research is based on English, most applications that build off that research are also tailored to English. Our approach to classifying…
Unfortunately, this is a very common occurrence in NLP applications. Our first step to combat this is through performing a spellcheck step when preprocessing all of our data. Next, some of the algorithms we employ only…
Our criteria mainly depended on internal testing to see which pre-packaged algorithms performed better on our data. While performance does vary with the domain (or in our case, industry of feedback we are analyzing), we…
To expand a little on the previous answer, existing cloud API's typically fall into one of the following categories: 1) generic NLP classification API's where you need your own large supply of training data and have…
Making the calls to the corresponding model from Flask was actually easier on AWS, since they were loaded into memory. Unfortunately, the scaling issues/excessive load times were big enough of an issue that we had to…
It probably would - however we built this out with a team of 1-2 engineers, so paying a bit more and dealing with maybe 20 seconds of extra latency a day is a price we are willing to pay for expedited/easier deployment.…
A bit of both. Flask was obviously not designed with serving Tensorflow models locally in mind, which is how we had it set up in v0. Towards the beginning we had to debug some weird threading issues, but towards the end…
Not inherently, but how we're using it makes it cheaper. Our stack has ~50 different ML models being served live, and GCP makes it easy to treat each model as a micro-service, and give auto-scaling to each one. This is…
While AWS does allow more hands-on control of the model stack, our "Custom Models" are more just different sets of weights using the exact same methodology. Basically, each customer that creates their own custom model…
For fine tuning BERT onto a specific domain, what amount of text data would you recommend to train on?
That depends on what you mean by noise. In the case of text classification, a lot of the noise in training is disagreement between human labelings. Unfortunately, classification systems will only be as good as the…
That is a good point. It helps to think of Precision and Recall (so by extension F1-Score) from your test data as random variables sampled from a distribution modeling the probability of getting each value in your…
I would imagine that AWS Comprehend utilizes a lot of customer feedback when training its internal models. However, they do advertise Comprehend as a general purpose NLP service, so they may have taken steps to make the…
Transfer learning is the premise of learning from data in one domain, and then being able to transfer that knowledge to another domain. In this case, we leverage large existing text datasets to build an initial model,…
There is definitely a feedback loop at play in the NLP community. Since most NLP research is based on English, most applications that build off that research are also tailored to English. Our approach to classifying…
Unfortunately, this is a very common occurrence in NLP applications. Our first step to combat this is through performing a spellcheck step when preprocessing all of our data. Next, some of the algorithms we employ only…
Our criteria mainly depended on internal testing to see which pre-packaged algorithms performed better on our data. While performance does vary with the domain (or in our case, industry of feedback we are analyzing), we…
To expand a little on the previous answer, existing cloud API's typically fall into one of the following categories: 1) generic NLP classification API's where you need your own large supply of training data and have…