This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
#!/usr/bin/env python | |
""" | |
Based on https://colab.research.google.com/drive/1mypqbHDrusZaIbqPoiEGY-WIbnpMHa2I?usp=sharing#scrollTo=bY0AwDiSxdVE | |
- Random Sampling | |
- After sorting (minimum padding but no randomness) | |
- Org Dynamic Batching (the one in speechbrain current version) | |
- **Mdf Dynamic Batching w/ fitted lognorm** (bucket boundaries set up with lognormal distribution fitted on dataset) | |
- **Mdf Dynamic Batching w/ fitted beta** (bucket boundaries set up with beta distribution fitted on dataset, mentioned in tuto) |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
#!/usr/bin/env python | |
import logging | |
import time | |
from typing import List, Optional | |
import numpy as np | |
import scipy.stats | |
import speechbrain | |
import torch |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
#!/usr/bin/env python | |
# Copyright 2022 Bofeng Huang | |
# coding=utf-8 | |
""" | |
Usage: | |
./scripts/convert_whisper_to_openai.py \ | |
--hf_model_name_or_path outputs/general/whisper-large-v2-ft-french-lr4e6-bs256-augment \ | |
--whisper_state_path outputs/general/whisper-large-v2-ft-french-lr4e6-bs256-augment/checkpoint_openai.pt | |
""" |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
{ | |
"name": "Vigogne Chat V3", | |
"inference_params": { | |
"input_prefix": "[INST]", | |
"input_suffix": "[/INST]", | |
"antiprompt": [ | |
"[INST]" | |
], | |
"pre_prompt": "[INST]<<SYS>>\\nVous êtes Vigogne, un assistant IA créé par Zaion Lab. Vous suivez extrêmement bien les instructions. Aidez autant que vous le pouvez.\\n<</SYS>>\\n\\n[/INST]" | |
} |