SupreethRao99

Moving to a specific GPU

cuda = torch.device('cuda')     # Default CUDA device, In our case it's Nvidia RTX 3090 
cuda0 = torch.device('cuda:0')  # specifies the CUDA device to be NVIDIA GeForce RTX 3090
cuda1 = torch.device('cuda:1')  # specifies the CUDA device to be NVIDIA GeForce RTX 3080

Example:
  x = torch.tensor([1., 2.], device=cuda0)
  # x.device is device(type='cuda', index=0)

SupreethRao99

some very long text afowngewngoewtg2w,
agnuwrgnuwrgwrgwrhwnu. wgnwoguwe w  hgr3g,
wgorgunrugnm3w[wrgw,
ewgwignwginweiuwengwingien,
ewrgiuewrbgoyiweynvn,
wqgwqgnin gq3-9nugq-gqg 3g93ugn2q-,
giqwnbbwiopngaowgnqwgqoghqprubq,
eboginwerobwienqq,

SupreethRao99

1. To convert to tflite, clone the TensorflowASR repository and build its dependencies

git clone https://github.com/TensorSpeech/TensorFlowASR.git
cd TensorFlowASR
pip3 install -e ".[tf2.8]"

2. Download the model weights and config file from here (https://drive.google.com/drive/folders/1rYpiYF0F9JIsAKN2DCFFtEdfNzVbBLHe?usp=sharing)
3. Add the downloaded config file to the same directory as the conversion file. ie to ./TensorFlowASR/examples/rnn_transducer/inference/
4. Change the code in gen_tflite_model.py to the code below

SupreethRao99

import torch
import torch.nn as nn
import torch.nn.functional as F

from typing import Optional

class WeightedKappaLoss(nn.Module):
  """
  Implements Weighted Kappa Loss. Weighted Kappa Loss was introduced in the
  [Weighted kappa loss function for multi-class classification

SupreethRao99