Struct rust_bert::pipelines::summarization::SummarizationConfig [−][src]
pub struct SummarizationConfig {}Show fields
pub model_type: ModelType, pub model_resource: Resource, pub config_resource: Resource, pub vocab_resource: Resource, pub merges_resource: Resource, pub min_length: i64, pub max_length: i64, pub do_sample: bool, pub early_stopping: bool, pub num_beams: i64, pub temperature: f64, pub top_k: i64, pub top_p: f64, pub repetition_penalty: f64, pub length_penalty: f64, pub no_repeat_ngram_size: i64, pub num_return_sequences: i64, pub num_beam_groups: Option<i64>, pub diversity_penalty: Option<f64>, pub device: Device,
Expand description
Configuration for text summarization
Contains information regarding the model to load, mirrors the GenerationConfig, with a different set of default parameters and sets the device to place the model on.
Fields
model_type: ModelType
Expand description
Model type
model_resource: Resource
Expand description
Model weights resource (default: pretrained BART model on CNN-DM)
config_resource: Resource
Expand description
Config resource (default: pretrained BART model on CNN-DM)
vocab_resource: Resource
Expand description
Vocab resource (default: pretrained BART model on CNN-DM)
merges_resource: Resource
Expand description
Merges resource (default: pretrained BART model on CNN-DM)
min_length: i64
Expand description
Minimum sequence length (default: 0)
max_length: i64
Expand description
Maximum sequence length (default: 20)
do_sample: bool
Expand description
Sampling flag. If true, will perform top-k and/or nucleus sampling on generated tokens, otherwise greedy (deterministic) decoding (default: true)
early_stopping: bool
Expand description
Early stopping flag indicating if the beam search should stop as soon as num_beam
hypotheses have been generated (default: false)
num_beams: i64
Expand description
Number of beams for beam search (default: 5)
temperature: f64
Expand description
Temperature setting. Values higher than 1 will improve originality at the risk of reducing relevance (default: 1.0)
top_k: i64
Expand description
Top_k values for sampling tokens. Value higher than 0 will enable the feature (default: 0)
top_p: f64
Expand description
Top_p value for Nucleus sampling, Holtzman et al.. Keep top tokens until cumulative probability reaches top_p (default: 0.9)
repetition_penalty: f64
Expand description
Repetition penalty (mostly useful for CTRL decoders). Values higher than 1 will penalize tokens that have been already generated. (default: 1.0)
length_penalty: f64
Expand description
Exponential penalty based on the length of the hypotheses generated (default: 1.0)
no_repeat_ngram_size: i64
Expand description
Number of allowed repetitions of n-grams. Values higher than 0 turn on this feature (default: 3)
num_return_sequences: i64
Expand description
Number of sequences to return for each prompt text (default: 1)
num_beam_groups: Option<i64>
Expand description
Number of beam groups for diverse beam generation. If provided and higher than 1, will split the beams into beam subgroups leading to more diverse generation.
diversity_penalty: Option<f64>
Expand description
Diversity penalty for diverse beam search. High values will enforce more difference between beam groups (default: 5.5)
device: Device
Expand description
Device to place the model on (default: CUDA/GPU when available)
Implementations
impl SummarizationConfig
[src]
impl SummarizationConfig
[src]pub fn new(
model_type: ModelType,
model_resource: Resource,
config_resource: Resource,
vocab_resource: Resource,
merges_resource: Resource
) -> SummarizationConfig
[src]
pub fn new(
model_type: ModelType,
model_resource: Resource,
config_resource: Resource,
vocab_resource: Resource,
merges_resource: Resource
) -> SummarizationConfig
[src]Instantiate a new summarization configuration of the supplied type.
Arguments
model_type
-ModelType
indicating the model type to load (must match with the actual data to be loaded!)- model_resource - The
Resource
pointing to the model to load (e.g. model.ot) - config_resource - The `Resource’ pointing to the model configuration to load (e.g. config.json)
- vocab_resource - The `Resource’ pointing to the tokenizer’s vocabulary to load (e.g. vocab.txt/vocab.json)
- merges_resource - The
Resource
pointing to the tokenizer’s merge file or SentencePiece model to load (e.g. merges.txt).
Trait Implementations
impl Default for SummarizationConfig
[src]
impl Default for SummarizationConfig
[src]fn default() -> SummarizationConfig
[src]
fn default() -> SummarizationConfig
[src]Returns the “default value” for a type. Read more
impl From<SummarizationConfig> for GenerateConfig
[src]
impl From<SummarizationConfig> for GenerateConfig
[src]fn from(config: SummarizationConfig) -> GenerateConfig
[src]
fn from(config: SummarizationConfig) -> GenerateConfig
[src]Performs the conversion.
Auto Trait Implementations
impl RefUnwindSafe for SummarizationConfig
impl Send for SummarizationConfig
impl Sync for SummarizationConfig
impl Unpin for SummarizationConfig
impl UnwindSafe for SummarizationConfig
Blanket Implementations
impl<T> BorrowMut<T> for T where
T: ?Sized,
[src]
impl<T> BorrowMut<T> for T where
T: ?Sized,
[src]pub fn borrow_mut(&mut self) -> &mut T
[src]
pub fn borrow_mut(&mut self) -> &mut T
[src]Mutably borrows from an owned value. Read more
impl<T> Instrument for T
[src]
impl<T> Instrument for T
[src]fn instrument(self, span: Span) -> Instrumented<Self>
[src]
fn instrument(self, span: Span) -> Instrumented<Self>
[src]Instruments this type with the provided Span
, returning an
Instrumented
wrapper. Read more
fn in_current_span(self) -> Instrumented<Self>
[src]
fn in_current_span(self) -> Instrumented<Self>
[src]impl<T> Pointable for T
impl<T> Pointable for T
impl<T> Same<T> for T
impl<T> Same<T> for T
type Output = T
type Output = T
Should always be Self
impl<V, T> VZip<V> for T where
V: MultiLane<T>,
impl<V, T> VZip<V> for T where
V: MultiLane<T>,