graffitist
graffitist copied to clipboard
Xilinx
Graph Transforms to Quantize and Retrain Deep Neural Nets in TensorFlow
Hello! I've found a performance issue in /scripts/imagenet_utils.py: `batch()` should be called before `map()`, which could make your program more efficient. Here is [the tensorflow document](https://tensorflow.google.cn/guide/data_performance?hl=zh_cn#vectorized_mapping) to support it. Detailed...
Dear Graffitist authors, I was looking through the provided models and recipes and was wondering if Yolo_v2_tiny trained thresholds and weights with INT4 weights was supported. Would it be possible...
Hi Graffitist Authors, Thanks for your great works. I just try to test your graffitist code: static and retrain mode. They both encouter the issues: 'UnboundLocalError: local variable 'input_patterns_list' referenced...
Hi there, thanks for the great work first. And I came across some strange loss during training, I just want to make sure the training procedure is the same as...
Hi, thanks again for sharing this repo for reproducing the awesome results. I am curious about is the BatchNorm layers folded into preceding Conv or FC layers while inference? I...
I was wondering if the real source code of graffitist is going to be released in future versions, rather than the byte-compiled python code and shared object files. This would...
Hi, Can we use graffitis to re-train a model quantized by Vitis-AI? BR,
In my network, there is a **depth_to_sparse** operation, which contains no learn-able parameters. What I want is to skip this to do a quantization aware training with this layer NOT...
Thanks for your work on this project and the paper "Trained Quantization Thresholds for Accurate and Efficient Fixed-Point Inference of Deep Neural Networks". Thesedays, I also want to use uniform...
