Maria Sargsyan

Turning off back-to-back optimizer does not disable fusing batch normalization layers into convolutional layers

6

**Describe the bug** Hi, I was converting CenterNet([CenterNet HourGlass104 512x512](centernet_hg104_512x512_coco17_tpu-8) from Tensorflow Object Detection API(https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/tf2_detection_zoo.md) with the Back-to-back optimizer turned off to disable the batchnorm fusion into conv layers following...

Hi, I am trying to understand the preprocessing of the EfficientnetV2. As far as I have understood except for resizing/cropping, the images are brought to the range [-1,1], and usually,...

Hi, I have been trying to export the v2 model to onnx and it fails. Do you know how to fix this? ``` tf.keras.backend.clear_session() model = effnetv2_model.EffNetV2Model(model_name=MODEL) _ = model(tf.ones([1,...

Hi, I have noticed that the Generator uses "track_running_stats= True". Is there a particular reason for it? From what I can understand when the track_running_stats is True it acts as...

Converting to BGR in style_transfer_net.py

1

Hey, I was going through the code and I cant understand why you convert the inputs from RGB to BGR. Would be grateful for any help, Best

Cannot do inference from the saved model for the model centernet_mobilenetv2fpn_512x512_coco17_od

1

# Prerequisites Please answer the following questions for yourself before submitting an issue. - [x] I am using the latest TensorFlow Model Garden release and TensorFlow 2. - [x] I...

Hey, I have noticed that the code for [tokenization](https://github.com/openai/CLIP/tree/main/clip)/simple_tokenizer.py) is very similiar to Gpt-2 encoding except that the vocabulary contains tokens ending with "". What is the meaning of ""?...

Hi there, I have been experimenting with the CLIP tokenizer and have observed that the tokenizer produces identical outputs for the following cases: 1). "1233" and "12 33" 2) "'Medicare...

Hi, I can see the in the tf/ folder there are two onnx files model-f6b98070.onnx and model-small.onnx but their name does not match to an of the models mentioned in...