model-optimization

model-optimization copied to clipboard

Issue for the FAKE_QUANT result

4

Hi, For the example: ` a=tf.fake_quant_with_min_max_vars([-1.0,0.0,1.0], min=-1, max=1)` ` sess=tf.Session()` ` sess(a)` The result is array([**-0.9960785**, 0. , **1.0039215**], dtype=float32) But according to the source code "fake_quant_ops_functor.h": it define the:...

jasonchenPJ

Quantizing model at runtime during training results in non-learned quantization

1

Prior to filing: check that this should be a bug instead of a feature request. Everything supported, including the compatible versions of TensorFlow, is listed in the overview page of...

FSet89

Tensorflow1.15 and TensorFlow2.4 were used for int8 precision quantization of the same Pb model, and the quantization results were different. Where, the BIAS of float32 bit is the same, but the weight of int8 bit convolution kernel is different

4

### 1. System information tensorflow1.15 、tensorflow2.4 ### 2. Code ``` converter = tf.lite.TFLiteConverter.from_saved_model(path) converter.post_training_quantize = True converter.target_ops = [tf.lite.OpsSet.TFLITE_BUILTINS, tf.lite.OpsSet.SELECT_TF_OPS] tflite_model = converter.convert() ```

carryzhang123

Pruning: improve custom training loop API

5

The recommended path for pruning with a custom training loop is not as simple as it could be. ``` pruned_model = setup_pruned_model() loss = tf.keras.losses.categorical_crossentropy optimizer = keras.optimizers.Adam() log_dir =...

alanchiao

Integer tflite model slow than float32 model

11

### 1. System information - quantized integer tflite model in Window10(without Nvidia GPU) - TensorFlow installation method pip installed tensorflow-gpu-2.2.2 and tensorflow-2.2.2. - try to run quantized model in mobile...

yahuuu

Can't run prunning with XNNPack

1

I am having the following error trying to prune an Efficientnet B1 network, but I am unable to understand even what it is referring to: Could not find `Conv2D 3x3`...

jhelsas

How to prune a custom tensor? The tensor is a recursive variable and is initialized with tf.zeros.

3

Prior to filing: check that this should be a bug instead of a feature request. Everything supported, including the compatible versions of TensorFlow, is listed in the overview page of...

starsky68

Fix the integration of TFLite models with multiple signatures and quantization

1

**System information** - TensorFlow version (you are using): tf-nightly - Are you willing to contribute it (Yes/No): No, this has a broader impact in the TFLite conversion & quantization API...

leandro-gracia-gil

Segmentation fault when invoking TFLite interpreter on basic quantized model

1

### 1. System information Operating System: Ubuntu 18.04.5 LTS Kernel: Linux 5.4.0-60-generic Architecture: x86-64 GPU: 2x Nvidia Quadro RTX8000 cuda: v11.0 Tensorflow v2.4.0 (installed through pip) Also tested on TF...

DavidvSon1

Model with SeparableConvs not converting to QAT

4

Prior to filing: check that this should be a bug instead of a feature request. Everything supported, including the compatible versions of TensorFlow, is listed in the overview page of...

dash-myelin