ContrastiveDataset iterator and ContrastiveDistillationDataset bug.

(updated)
## ContrastiveDataset

ContrastiveDataset by default generates and builds the whole positive and negative examples internally in lists. Then in trainer.py
https://github.com/huggingface/setfit/blob/146c7c9dacdc7dca678b2fffff8ddeb79dd762c2/src/setfit/trainer.py#L605
from_list method is used to create the dataset for training. This is fine with smaller datasets **which I recognize is the whole point of the setfit method**, yet if you want to build bigger dataset for whatever reason this can blow your RAM memory (as I have experienced). It would be best if you can build the dataset from a generator ```Dataset.from_generator``` While the current ```__iter__``` method is correct the internal data that it uses is eagerly generated on ```__init__``` . It would be best if the generator is truly "lazy".

Other question I am not sure why is what is the reason that replacement is set to True by default? Although there are some cases where this makes sense the matrix picture in the sampling guide shows the diagonal as empty (ie I wont expect the library to do the opposite by default - https://huggingface.co/docs/setfit/v1.1.0/en/conceptual_guides/sampling_strategies) Also there is no way to control that.

https://github.com/huggingface/setfit/blob/146c7c9dacdc7dca678b2fffff8ddeb79dd762c2/src/setfit/sampler.py#L15

Using this setting we get pairs such as:
("The movie was awesome", "The movie was awesome") as a positive pair in the training process.

## ContrastiveDistillationDataset

ContrastiveDistillationDataset inherits from ContrastiveDataset but set its own

https://github.com/huggingface/setfit/blob/146c7c9dacdc7dca678b2fffff8ddeb79dd762c2/src/setfit/sampler.py#L170

by looking at the code you would expect that the pairs are generated using these ```self.sentence_labels``` but what happens is the ```super().__init__``` goes first and there it uses the sentence_labels from ContrastiveDataset
https://github.com/huggingface/setfit/blob/146c7c9dacdc7dca678b2fffff8ddeb79dd762c2/src/setfit/sampler.py#L63
and this results in creating the pairs using the function defined in ContrastiveDistillationDataset
https://github.com/huggingface/setfit/blob/146c7c9dacdc7dca678b2fffff8ddeb79dd762c2/src/setfit/sampler.py#L178
but with the sentence_labels defined by ContrastiveDataset.
This results in creating pairs of sentences that are always selecting the first row first column element from the cos_matrix as label.
this is what you expect by looking at the code
```
{"sentence_1": text_one, "sentence_2": text_two, "label": self.cos_sim_matrix[id_one][id_two]}
```
But because the init method runs with the sentence_labels generated from the parent you always get:
```
{"sentence_1": text_one, "sentence_2": text_two, "label": self.cos_sim_matrix[0][0]}
```
because the labels in the parent are set as 0 for each sentence. Then after the init you set ```self.sentence_labels = list(enumerate(self.sentences))``` but this is never used as the pairs are already generated, very nasty and easily overlooked bug. This unfortunately makes ContrastiveDistillationDataset unusable.

Start in src/setfit/sampler.py with ContrastiveDataset and ContrastiveDistillationDataset constructors and pair-generation methods, then inspect the Dataset.from_list call in src/setfit/trainer.py. Reproduce the eager memory use and the distillation labels, and consider the replacement behavior described in the issue. Done means large datasets are not eagerly materialized and distillation pairs use the intended labels, with regression coverage for the reported cases.

Written by the indexing model from the issue text.