tinyml
tinyml copied to clipboard
mit-han-lab
Hi, I recognized that we have ckpt and json files there and I`m trying to transform the module into onnx format but I can`t find the corresponding neural network file...
What if NetAug is applied to the tiny models searched from once-for-all, since once-for-all is trained in a similar way? Will NetAug boost the tiny ofa models?
Hi, thanks for the great work! We implemented the update-bias-only transfer learning in our own codebase, however, we didn't notice a huge decrease of peak memory / memory usage during...
``` torchpack dist-run -np 32 -H $server1:8, $server2:8, $server3:8, $server4:8 \ python train.py configs/netaug.yaml \ --data_provider "{/mnt/dolphinfs/hdd_pool/docker/user/hadoop-basecv/maxin/data/ILSVRC2012, image_size:176, base_batch_size:64}" \ --run_config "{base_lr:0.0125}" \ --model "{name:mcunet}" \ --netaug "{aug_expand_list:[1.0,1.6,2.2,2.8], aug_width_mult_list:[1.0,1.6,2.2]}" \...
Hi, While exporting the tiny model from a large model, I found there was an accuracy drop. Can you clarify this part? Is there anything that needs to do before...
The pretrained models you located in Drive do not have .pt files. Also, if we change the name of the init file to .pt file, the accuracy of the model...
