| models: | |
| # AWQ Int4 sẵn (~16GB VRAM) — vừa Colab L4 24GB. Bản full + bitsandbytes 4-bit cần >24GB. | |
| omni: "cyankiwi/Qwen3-Omni-30B-A3B-Instruct-AWQ-4bit" | |
| quant: "awq" # awq | 4bit (bitsandbytes, cần GPU lớn hơn) | |
| diarize: "pyannote/speaker-diarization-community-1" | |
| translate: | |
| source_lang: "Chinese" | |
| target_lang: "Vietnamese" | |
| chars_per_sec: 22 | |
| max_line_chars: 52 | |
| max_lines: 2 | |
| correct_ocr: false # Bước 1 — OCR sửa phụ đề cháy hình (mặc định tắt) | |
| output_suffix: ".vi.srt" # giữ nếp cũ | |
| scene: | |
| max_gap: 1.5 | |
| max_cues: 4 | |
| max_dur: 20.0 | |
| profiling: | |
| clips_per_speaker: 5 | |
| max_seconds_per_clip: 8 | |
| sampling: | |
| temperature: 0.7 | |
| top_p: 0.95 | |
| top_k: 64 | |
| max_new_tokens: 1024 | |
| paths: | |
| drive_root: "/content/drive/MyDrive/Gemma" | |