How to use the @tensorflow/tfjs-node.mul function in @tensorflow/tfjs-node

this.criticOptimizer.minimize(() => {
        // Compute Q targets for current states (y-i)
        const qTargets = tf.add(rewards, tf.mul(tf.mul(gamma, qTargetsNext), tf.sub(1, dones)));
        const qExpected = this.critic.predict([states, actions]);
        const criticLoss = tf.losses.meanSquaredError(qExpected, qTargets);
        // torch.nn.utils.clip_grad_norm_(self.critic.parameters(), 1)
        return criticLoss;
      });
      // Actor update

const res = sWeights.map((s, i) => 
        tf.mul(s, tau).add(tf.mul(tWeights[i], 1-tau)));
      layer.setWeights(res);

const action = tf.tidy(() => {
      let action = tf.squeeze(this.actor.predict(tf.tensor([state])));
      if (train) {
        const noise = softmax(this.noise.sample());        
        action = action.mul(1-this.epsilon).add(tf.mul(noise, this.epsilon));
      }
      return action;
    });
    const data = await action.data();