Kevin Chih Yao Huang

icon company Notable Health icon location Head of Data and Machine Learning at Notable Health

Results 14 comments of Kevin Chih Yao Huang

exuse me！！！！i have question,my data is image which full of long words,just like books,but when i train model by my data,it can not detect all the words in one line.maybe it can only detect half of a line and my input_size is 768 and text_scale is 768.could you tell me where i maybe wrong?thank you so much

What sort of aspect ratio of "long text" are you looking to detect? The two main knobs for optimizing performance on long text is: 1) resolution (essentially number of pixels...

Question about training with empty labels

> Hello everyone, > > Is it possible to explicitly train the model to recognize 'empty' images? > > I have a use-case where the speed of the model is...

Can't pip3 install kenlm on Ubuntu Focal

> can someone please help me figure out step 4 pip install from the checkout, what is the checkout? just go to the kenlm directory (where setup.py lives) and run...

TF 2 Implementation error

Have you tried `tf.raw_ops.MutableHashTable`? https://www.tensorflow.org/api_docs/python/tf/raw_ops/MutableHashTable The interface seems like it should do the same as `tf.contrib.lookup.MutableHashTable`

Tensorflow 2 compatibility

@emedvedev I'd be happy to contribute to a TF2 conversion. A branch for folks to merge into would be very useful.

Tensorflow 2 compatibility

Awesome, by the way, this is to completely convert to TF2 right? Not using tensorflow.compat.v1 Let me know and I'll take a stab at it.

Tensorflow 2 compatibility

> @OctaM Yes that would be really helpful. > By the way i have converted most of the part using tf compat v1 and I have stuck when there is...

Tensorflow 2 compatibility

> I performed the same steps and got stuck there as well. The main issue seems to be the fact that the project is very dependent on the `contrib` module...

Tensorflow 2 compatibility

@emedvedev I've read through the code, doing some planning now in terms of the code structuring. Overall, would you mind if we convert to tf.keras API where possible? In TF2...

Streaming Responses As Ouput Using FastAPI Support

> Thank you for your response! Unfortunately, it throws this error: `TypeError: 'async for' requires an object with __aiter__ method, got str` to this line: `async for message in await...