Reads in a token and returns booleans that indicate the progress made by it. This function will update the state of this object unlik

Reads in a token and returns booleans that indicate the progress made by it. This function will update the state of this object unlik

r""" Save a generation configuration object to the directory `save_directory`, so that it can be re-loaded using the [`~GenerationConf

Creates a new instance of this constraint. Args: stateful(`bool`): Whether to not only copy the constraint for new insta

Add a new hypothesis to the list.

Creates a new instance of this constraint. Args: stateful(`bool`): Whether to not only copy the constraint for new insta

Function that is called by `.generate()` to push new tokens

r""" Save a generation configuration object to the directory `save_directory`, so that it can be re-loaded using the [`~GenerationConf

Add a new hypothesis to the list.

Serializes this instance to a Python dictionary. Returns: `Dict[str, Any]`: Dictionary of all the attributes that make u

Gather value of `tensors` (tensor or list/tuple of nested tensors) and convert them to numpy before concatenating them to `gathered`

r""" Instantiate a [`GenerationConfig`] from a generation configuration file. Args: pretrained_model_name (`str` or `os.P

Whether or not this process is the global main process (when training in a distributed fashion on several machines, this is only goin

Instantiates a [`GenerationConfig`] from a Python dictionary of parameters. Args: config_dict (`Dict[str, Any]`):

Instantiates a [`GenerationConfig`] from a Python dictionary of parameters. Args: config_dict (`Dict[str, Any]`):

Log `logs` on the various objects watching training. Subclass and override this method to inject custom behavior. Args:

Function that is called by `.generate()` to push new tokens

Serializes this instance to a Python dictionary. Returns: `Dict[str, Any]`: Dictionary of all the attributes that make u

Expands tensors from [batch_size, ...] to [batch_size * expand_size, ...]

Expands tensors from [batch_size, ...] to [batch_size * expand_size, ...]

Expands tensors from [batch_size, ...] to [batch_size * expand_size, ...]

Gathers the beam slices indexed by beam_indices into new beam array.

Gathers the beam slices indexed by beam_indices into new beam array.

Function that is called by `.generate()` to signal the end of generation

This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch, num_key_value_heads, seqlen, he

This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch, num_key_value_heads, seqlen, he

This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch, num_key_value_heads, seqlen, he

This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch, num_key_value_heads, seqlen, he

This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch, num_key_value_heads, seqlen, he

Given the (decoder/cross attentions)/(decoder hidden states) for multiple generated tokens, splits it into a tuple where each member correspo

Applies Rotary Position Embedding to the query and key tensors. Args: q (`torch.Tensor`): The query tensor. k (`torch.Tensor`): T

Applies Rotary Position Embedding to the query and key tensors. Args: q (`torch.Tensor`): The query tensor. k (`torch.Tensor`): T

If there are enough hypotheses and that none of the hypotheses being generated can become better than the worst one in the heap, then

Helper to get number of samples in a [`~torch.utils.data.DataLoader`] by accessing its dataset. When dataloader.dataset does not exis

Upload `self.model` and `self.tokenizer` to the 🤗 model hub on the repo `self.args.hub_model_id`. Parameters: commit_mes

Resets the state of this constraint to its initialization. We would call this in cases where the fulfillment of a constraint is abrup

Resets the state of this constraint to its initialization. We would call this in cases where the fulfillment of a constraint is abrup

Serializes this instance to a JSON string. Args: use_diff (`bool`, *optional*, defaults to `True`): If s

Serializes this instance to a JSON string. Args: use_diff (`bool`, *optional*, defaults to `True`): If s

Validates the values of the attributes of the [`GenerationConfig`] instance. Raises exceptions in the presence of parameterization th

Standardizes the format of the cache so as to match most implementations, i.e. to tuple(tuple([batch_size, num_heads, ...]))

Expands tensors from [batch_size, ...] to [batch_size * expand_size, ...] or [batch_size, expand_size, ...], depending on `expand_in_

Expands tensors from [batch_size, ...] to [batch_size * expand_size, ...] or [batch_size, expand_size, ...], depending on `expand_in_

Implement in subclasses of [`PreTrainedModel`] for custom behavior to adjust the logits in the generate method.

Implement in subclasses of [`PreTrainedModel`] for custom behavior to adjust the logits in the generate method.

When called, returns the token that would take this constraint one step closer to being fulfilled. Return: token_ids(`to

When called, returns the token that would take this constraint one step closer to being fulfilled. Return: token_ids(`to

Function that is called by `.generate()` to signal the end of generation

Instantiates a [`GenerationConfig`] from a [`PretrainedConfig`]. This function is useful to convert legacy [`PretrainedConfig`] objec

r""" Instantiate a [`GenerationConfig`] from a generation configuration file. Args: pretrained_model_name (`str` or `os.P

Whether or not this process is the local (e.g., on one machine if training in a distributed fashion on several machines) main process

This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch, num_key_value_heads, seqlen, he

This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch, num_key_value_heads, seqlen, he

This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch, num_key_value_heads, seqlen, he

Main training entry point. Args: resume_from_checkpoint (`str` or `bool`, *optional*): If a `str`, local

Validates the values of the attributes of the [`GenerationConfig`] instance. Raises exceptions in the presence of parameterization th