kumariko comments

Results 6 comments of


                                            kumariko

Reshape layer drops mask from previous layers

@jvishnuvardhan I Was able to reproduce the issue in TF **[v2.6](https://colab.research.google.com/gist/kumariko/c5dd4b87a86cec75578028f1c202e2d0/15194.ipynb#scrollTo=8TpXmySqd-YQ)** and **[TF-nightly](https://colab.research.google.com/gist/kumariko/1ea7e3a7a39fb5e1334a75b5fdf41221/15194.ipynb#scrollTo=aJJtjFqWI9cf)** please find the gist .Thanks!

Yolov5 V6 The exported TFJS model is not available because img SIez cannot be set to 300

@gamefake588 Could you please have a look on the [link1](https://github.com/ultralytics/yolov5/issues/2127), [link2](https://github.com/tensorflow/tfjs/issues/2046) and let us know if it helps? Thanks!

iterating over `tf.Tensor` is not allowed: AutoGraph did convert this function. This might indicate you are trying to use an unsupported feature

@827346462 Please refer to this [**`link`** ](https://stackoverflow.com/questions/65059475/iterating-over-tf-tensor-is-not-allowed-autograph-did-convert-this-function)and let us know if it helps? In order to expedite the trouble-shooting process, please provide the Code snippet which you are using. Thanks!

Updating Documentation link in export_inference_graph.py

@tombstone @pkulzc @jch1 any update on this PR?

Tokenizer converts padding integers to OOV when oov_token is not None

@ymodak Was able to reproduce the issue in TF v2.6 , TF v2.5 , TF-nightly please find the gist [**`here`**](https://colab.research.google.com/gist/kumariko/4dde948009a449a7b0ace9011ca14874/51376.ipynb#scrollTo=_MJn1e6qlg1B).Thanks!

Couldn't convert ssd_efficientdet_d0_512x512_coco17_tpu-8 model to int8 tflite model

@xiang-burlington Could you please have a look on the [link1](https://stackoverflow.com/questions/66767195/run-inference-with-quantized-tflite-model-int8-in-python) , [link2](https://stackoverflow.com/questions/67330342/converting-ssd-object-detection-model-to-tflite-and-quantize-it-from-float-to-ui), [link3](https://stackoverflow.com/questions/66551794/tflite-cant-quantize-the-input-and-output-of-tensorflow-model-to-int8), [document](https://www.tensorflow.org/lite/convert) and let us know if it helps? Thanks!