whisper_full_params class final

Parameters for the whisper_full() function If you change the order or add new parameters, make sure to update the default values in whisper.cpp: whisper_full_default_params()

Inheritance
Implemented types
Available extensions

Constructors

whisper_full_params.new()

Properties

abort_callback ↔ int
called each time before ggml computation starts
getter/setter pair
abort_callback_user_data ↔ Pointer<Void>
getter/setter pair
address → Pointer<T>

Available on T, provided by the StructAddress extension

The memory address of the underlying data.
no setter
audio_ctx ↔ int
overwrite the audio context size (0 = use default)
getter/setter pair
getter/setter pair
debug_mode ↔ bool
enable debug_mode provides extra info (eg. Dump log_mel)
getter/setter pair
detect_language ↔ bool
getter/setter pair
duration_ms ↔ int
audio duration to process in ms
getter/setter pair
encoder_begin_callback ↔ whisper_encoder_begin_callback
called each time before the encoder starts
getter/setter pair
encoder_begin_callback_user_data ↔ Pointer<Void>
getter/setter pair
entropy_thold ↔ double
similar to OpenAI's "compression_ratio_threshold"
getter/setter pair
grammar_penalty ↔ double
getter/setter pair
grammar_rules ↔ Pointer<Pointer<whisper_grammar_element>>
getter/setter pair
greedy ↔ UnnamedStruct1
getter/setter pair
hashCode → int
The hash code for this object.
no setterinherited
i_start_rule ↔ int
getter/setter pair
initial_prompt ↔ Pointer<Char>
tokens to provide to the whisper decoder as initial prompt these are prepended to any existing text context from a previous call use whisper_tokenize() to convert text to tokens maximum of whisper_n_text_ctx()/2 tokens are used (typically 224)
getter/setter pair
language ↔ Pointer<Char>
for auto-detection, set to nullptr, "" or "auto"
getter/setter pair
length_penalty ↔ double
ref: https://github.com/openai/whisper/blob/f82bc59f5ea234d4b97fb2860842ed38519f7e65/whisper/transcribe.py#L267
getter/setter pair
logits_filter_callback ↔ whisper_logits_filter_callback
called by each decoder to filter obtained logits
getter/setter pair
logits_filter_callback_user_data ↔ Pointer<Void>
getter/setter pair
logprob_thold ↔ double
getter/setter pair
max_initial_ts ↔ double
ref: https://github.com/openai/whisper/blob/f82bc59f5ea234d4b97fb2860842ed38519f7e65/whisper/decoding.py#L97
getter/setter pair
max_len ↔ int
max segment length in characters
getter/setter pair
max_tokens ↔ int
max tokens per segment (0 = no limit)
getter/setter pair
n_grammar_rules ↔ int
getter/setter pair
n_max_text_ctx ↔ int
max tokens to use from past text as prompt for the decoder
getter/setter pair
n_threads ↔ int
getter/setter pair
new_segment_callback ↔ whisper_new_segment_callback
called for every newly generated text segment
getter/setter pair
new_segment_callback_user_data ↔ Pointer<Void>
getter/setter pair
no_context ↔ bool
do not use past transcription (if any) as initial prompt for the decoder
getter/setter pair
no_speech_thold ↔ double
getter/setter pair
no_timestamps ↔ bool
do not generate timestamps
getter/setter pair
offset_ms ↔ int
start offset in ms
getter/setter pair
print progress information
getter/setter pair
print results from within whisper.cpp (avoid it, use callback instead)
getter/setter pair
print special tokens (e.g.
getter/setter pair
print timestamps for each text segment when printing realtime
getter/setter pair
progress_callback ↔ whisper_progress_callback
called on each progress update
getter/setter pair
progress_callback_user_data ↔ Pointer<Void>
getter/setter pair
prompt_n_tokens ↔ int
getter/setter pair
prompt_tokens ↔ Pointer<whisper_token>
getter/setter pair
runtimeType → Type
A representation of the runtime type of the object.
no setterinherited
single_segment ↔ bool
force single segment output (useful for streaming)
getter/setter pair
split_on_word ↔ bool
split on word rather than on token (when used with max_len)
getter/setter pair
strategy ↔ int
getter/setter pair
suppress_blank ↔ bool
ref: https://github.com/openai/whisper/blob/f82bc59f5ea234d4b97fb2860842ed38519f7e65/whisper/decoding.py#L89
getter/setter pair
suppress_nst ↔ bool
non-speech tokens, ref: https://github.com/openai/whisper/blob/7858aa9c08d98f75575035ecd6481f462d66ca27/whisper/tokenizer.py#L224-L253
getter/setter pair
suppress_regex ↔ Pointer<Char>
A regular expression that matches tokens to suppress
getter/setter pair
tdrz_enable ↔ bool
enable tinydiarize speaker turn detection
getter/setter pair
temperature ↔ double
initial decoding temperature, ref: https://ai.stackexchange.com/a/32478
getter/setter pair
temperature_inc ↔ double
fallback parameters ref: https://github.com/openai/whisper/blob/f82bc59f5ea234d4b97fb2860842ed38519f7e65/whisper/transcribe.py#L274-L278
getter/setter pair
thold_pt ↔ double
timestamp token probability threshold (~0.01)
getter/setter pair
thold_ptsum ↔ double
timestamp token sum probability threshold (~0.01)
getter/setter pair
token_timestamps ↔ bool
enable token-level timestamps
getter/setter pair
translate ↔ bool
getter/setter pair

Methods

noSuchMethod(Invocation invocation) → dynamic
Invoked when a nonexistent method or property is accessed.
inherited
toString() → String
A string representation of this object.
inherited

Operators

operator ==(Object other) → bool
The equality operator.
inherited