LatentQuantize exploding loss

I'm trying to use the LatentQuantize model in an autoencoder context. My inputs are flat 1-d tensors (32) and my encoder passes a shape of (batch_size, 64) to the quantizer. For now, my "levels" is [8, 6, 4], my latent_dim is 64:

```
self.lq = LatentQuantize(
levels=levels,
dim=latent_dim,
commitment_loss_weight=0.1,
quantization_loss_weight=0.1,
)
```

The loss starts at zero, then exponentially increases:

Any thoughts as to why this might happen?

Start with the LatentQuantize entry point and reproduce the reported autoencoder setup: flat 1-D inputs of 32, an encoder output of (batch_size, 64), levels [8, 6, 4], and latent_dim 64. Trace the loss from its initial zero value through the exponential increase, and document the conditions or configuration that produce the behavior.

Written by the indexing model from the issue text.