Shorter names for STT scripts (#30)

2025-12-25 20:39:56 +00:00 · 2025-07-02 18:21:42 +02:00
parent 07ac744609
commit 433dca3751
8 changed files with 12 additions and 22 deletions
--- a/scripts/stt_evaluate_on_dataset.py
+++ b/scripts/stt_evaluate_on_dataset.py
--- a/scripts/transcribe_from_file_via_pytorch.py
+++ b/scripts/transcribe_from_file_via_pytorch.py
--- a/scripts/transcribe_from_file_via_rust_server.py
+++ b/scripts/transcribe_from_file_via_rust_server.py
--- a/scripts/transcribe_from_file_via_pytorch_with_prompt.py
+++ b/scripts/transcribe_from_file_via_pytorch_with_prompt.py
@@ -14,15 +14,7 @@ import tqdm


 class PromptHook:
-    def __init__(
-        self,
-        tokenizer,
-        prefix,
-        padding_tokens=(
-            0,
-            3,
-        ),
-    ):
+    def __init__(self, tokenizer, prefix, padding_tokens=(0, 3)):
        self.tokenizer = tokenizer
        self.prefix_enforce = deque(self.tokenizer.encode(prefix))
        self.padding_tokens = padding_tokens
@@ -141,7 +133,7 @@ def main(args):
        prompt_frames = audio_prompt.shape[1] // mimi.frame_size
        no_prompt_offset_seconds = audio_delay_seconds + audio_silence_prefix_seconds
        no_prompt_offset = int(no_prompt_offset_seconds * mimi.frame_rate)
-        text_tokens = text_tokens[prompt_frames + no_prompt_offset:]
+        text_tokens = text_tokens[prompt_frames + no_prompt_offset :]

    text = tokenizer.decode(
        text_tokens[text_tokens > padding_token_id].numpy().tolist()
--- a/scripts/transcribe_from_mic_via_mlx.py
+++ b/scripts/transcribe_from_mic_via_mlx.py
--- a/scripts/transcribe_from_mic_via_rust_server.py
+++ b/scripts/transcribe_from_mic_via_rust_server.py