r""" Generates sequences of token ids for models with a language modeling head. <Tip warning={true}> Most generation-contro

r""" Generates sequences of token ids for models with a language modeling head. <Tip warning={true}> Most generation-contro

Returns vocab as a dict

Returns vocab as a dict

Initializes a git repo in `self.args.hub_model_id`.

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

Prints the new text to stdout. If the stream is ending, also prints a newline.

Prints the new text to stdout. If the stream is ending, also prints a newline.

Perform an evaluation step on `model` using `inputs`. Subclass and override to inject custom behavior. Args: mo

Sets values in the deepspeed plugin based on the Trainer args

Returns the number of remaining steps of `advance()` in order to complete this constraint.

Returns the number of remaining steps of `advance()` in order to complete this constraint.

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

Rotates half the hidden dims of the input.

Rotates half the hidden dims of the input.

Rotates half the hidden dims of the input.

Rotates half the hidden dims of the input.

Rotates half the hidden dims of the input.

Rotates half the hidden dims of the input.

Rotates half the hidden dims of the input.

Rotates half the hidden dims of the input.

Rotates half the hidden dims of the input.

Rotates half the hidden dims of the input.

r""" Generates sequences of token ids for models with a language modeling head using **multinomial sampling** and can be used for text

r""" Activates the neftune as presented in this code: https://github.com/neelsjain/NEFTune and paper: https://arxiv.org/abs/2310.05914

This function can be overwritten in the specific modeling_flax_<model-name>.py classes to allow for custom beam search behavior. Note

This function can be overwritten in the specific modeling_flax_<model-name>.py classes to allow for custom beam search behavior. Note

This beam search function is heavily inspired by Flax's official example: https://github.com/google/flax/blob/main/examples/wmt/decod

This beam search function is heavily inspired by Flax's official example: https://github.com/google/flax/blob/main/examples/wmt/decod

Copied from fairseq for no_repeat_ngram in beam_search

Copied from fairseq for no_repeat_ngram in beam_search

This function takes projected key, value states from a single input token and concatenates the states to cached states from previous

This function takes projected key, value states from a single input token and concatenates the states to cached states from previous

This function takes projected key, value states from a single input token and concatenates the states to cached states from previous

This function takes projected key, value states from a single input token and concatenates the states to cached states from previous

Converts an index (integer) in a token (str) using the vocab.

Converts an index (integer) in a token (str) using the vocab.

Crops the past key values up to a certain maximum length.

Crops the past key values up to a certain maximum length.

Deactivates the neftune method. Make sure to call `_activate_neftune` first.

dynamic RoPE layers should recompute `inv_freq` in the following situations: 1 - growing beyond the cached sequence length (allow sca

Expands attention_mask from `[bsz, seq_len]` to `[bsz, 1, tgt_seq_len, src_seq_len]`.

Expands attention_mask from `[batch_size, src_length]` to `[batch_size, 1, tgt_length, src_length]`.

Expands attention_mask from `[bsz, seq_len]` to `[bsz, 1, tgt_seq_len, src_seq_len]`.

Calls the forward method of Flash Attention - if the input hidden states contain at least one padding token first unpad the input, th

Calls the forward method of Flash Attention - if the input hidden states contain at least one padding token first unpad the input, th

Calls the forward method of Flash Attention - if the input hidden states contain at least one padding token first unpad the input, th

Calls the forward method of Flash Attention - if the input hidden states contain at least one padding token first unpad the input, th

Calls the forward method of Flash Attention - if the input hidden states contain at least one padding token first unpad the input, th

Returns the generation mode triggered by a [`GenerationConfig`] instance.

This class returns a [`LogitsProcessorList`] list object that contains all relevant [`LogitsProcessor`] instances used to modify the

This class returns a [`FlaxLogitsProcessorList`] list object that contains all relevant [`FlaxLogitsProcessor`] instances used to mod

This class returns a [`TFLogitsProcessorList`] list object that contains all relevant [`TFLogitsProcessor`] instances used to modify

This class returns a [`LogitsProcessorList`] list object that contains all relevant [`LogitsProcessor`] instances used to modify the

This class returns a [`LogitsProcessorList`] list object that contains all relevant [`LogitsProcessor`] instances used to modify the

This class returns a [`FlaxLogitsProcessorList`] list object that contains all relevant [`FlaxLogitsProcessor`] instances used to mod

This class returns a [`TFLogitsProcessorList`] list object that contains all relevant [`TFLogitsProcessor`] instances used to modify

This class returns a [`FlaxLogitsProcessorList`] list object that contains all relevant [`FlaxLogitsWarper`] instances used for multi

This class returns a [`FlaxLogitsProcessorList`] list object that contains all relevant [`FlaxLogitsWarper`] instances used for multi