IsraelHebrew speechWhisper Large v348K+ downloads

Explore Whisper Large v3 Hebrew

Transcribe Hebrew speech with ivrit.ai's fine-tuned Whisper Large v3 model, including optional segment or word timestamps. Use the playground to give Whisper Large v3 Hebrew a representative input and review its output. Adjust the available settings to shape the result for your task.

Model

ivrit-ai/whisper-large-v3

Input

Configure your request

Audio passed directly to AutomaticSpeechRecognitionPipeline inputs.

Documentation

Whisper Large v3 Hebrew field reference

Use this reference to understand how every field in the playground changes the request sent to ivrit-ai/whisper-large-v3.

inputs

Required

Type: File

Uploads the Hebrew speech recording to transcribe. Clear speech, limited background noise, and an appropriate sample rate improve recognition; long files may benefit from chunking and overlap controls.

return_timestamps

Optional

Type: String

Default: false

Values: false, true, word

Controls timestamp detail in the transcript. False returns text only, True returns segment-level timing, and Word requests timing for individual words; finer timestamps increase response detail and may be less precise on noisy audio.

chunk_length_s

Optional

Type: Number

Default: None

Range: 0–Unbounded

Splits long audio into chunks of this many seconds before transcription. Chunking controls memory use and enables long recordings, but chunks that are too short provide less linguistic context and may reduce accuracy.

stride_length_s

Optional

Type: Text

Default: None

Adds overlapping audio on the left and right of each chunk, supplied as one number or a [left, right] pair in seconds. Overlap reduces cut-off words at boundaries but increases duplicated processing time.

ignore_warning

Optional

Type: Boolean

Default: Off

Suppresses the pipeline warning shown when external chunking is used with Whisper’s own long-form behavior. It changes only warning output, not transcription; enable it only after confirming the chosen chunking approach is intentional.

return_language

Optional

Type: Boolean

Default: Off

Includes the detected or selected language in the transcription result. This is useful for auditing multilingual inputs, though this Hebrew checkpoint is best used with the language explicitly set to Hebrew in generation options.

batch_size

Optional

Type: Integer

Default: 1

Range: 1–Unbounded

Controls how many items are evaluated in parallel during one forward pass. Larger batches can improve throughput but use more memory; reduce the value if processing runs out of device memory.

generate_kwargs

Optional

Type: JSON object

Default: {"language":"he","task":"transcribe"}

Provides advanced keyword arguments directly to the model’s generation method. Use it for supported controls that are not shown separately in the form, and avoid duplicating named settings because these values may override or conflict with them.

decoder_kwargs

Optional

Type: JSON object

Default: None

Passes a JSON object of keyword arguments to the pipeline decoder during transcript post-processing. Use it for decoder-specific controls not exposed elsewhere, and leave it empty for the model defaults because unsupported keys can cause decoding errors.