Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,52 @@
|
|
| 1 |
---
|
| 2 |
-
license:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
license: other
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
+
tags:
|
| 6 |
+
- automatic-speech-recognition
|
| 7 |
+
- speech-recognition
|
| 8 |
+
- state-space-models
|
| 9 |
+
- mamba
|
| 10 |
+
- ctc
|
| 11 |
+
- librispeech
|
| 12 |
+
- pytorch
|
| 13 |
+
- masters-thesis
|
| 14 |
+
pretty_name: SSM-AST thesis checkpoints
|
| 15 |
---
|
| 16 |
+
|
| 17 |
+
# SSM-AST: State Space Models for Automatic Speech Transcription
|
| 18 |
+
|
| 19 |
+
This repository contains model artifacts for the master's thesis **βState Space Models for Automatic Speech Transcription.β** It provides the trained acoustic encoder checkpoints, language-model checkpoints, n-gram training text, and selected training logs used to support evaluation of a pure State Space Model (SSM) automatic speech transcription pipeline on LibriSpeech.
|
| 20 |
+
|
| 21 |
+
The code for training and evaluation is maintained separately. This Hugging Face repository is intended as a checkpoint and artifact archive so that the thesis evaluation pipeline can be run without retraining the full models from scratch.
|
| 22 |
+
|
| 23 |
+
## Repository contents
|
| 24 |
+
|
| 25 |
+
```text
|
| 26 |
+
SSM-AST/
|
| 27 |
+
βββ datasets/
|
| 28 |
+
β βββ librispeecm_lm_dataset_pre-processed_char_level_text.txt
|
| 29 |
+
βββ encoder_checkpoints/
|
| 30 |
+
β βββ enc_mamba3_460h_checkpoint_best_epoch=49_val_wer=0.255.ckpt
|
| 31 |
+
β βββ enc_mamba3_960h_checkpoint_best_epoch=91_val_wer=0.186.ckpt
|
| 32 |
+
β βββ enc_mamba_460h_checkpoint_best_epoch=49_val_wer=0.227.ckpt
|
| 33 |
+
β βββ enc_mamba_960h_checkpoint_best_epoch=100_val_wer=0.155.ckpt
|
| 34 |
+
β βββ enc_ssssm_460h_checkpoint_best_epoch=49_val_wer=0.197.ckpt
|
| 35 |
+
β βββ enc_ssssm_960h_checkpoint_best_epoch=100_val_wer=0.111.ckpt
|
| 36 |
+
β βββ enc_ssssm_960h_checkpoint_best_epoch=98_val_wer=0.111.ckpt
|
| 37 |
+
βββ lm_checkpoints/
|
| 38 |
+
β βββ lm_mamba3_checkpoint_MaxChars-1000000000_ds-64_d320_L18.pt
|
| 39 |
+
β βββ lm_mamba_checkpoint_MaxChars-1000000000_d320_L18.pt
|
| 40 |
+
β βββ lm_ngram_checkpoint_char_10gram.pkl
|
| 41 |
+
βββ log files/
|
| 42 |
+
β βββ 0utput_exp-mamba-1_960h_W-320_D-48_S-16_B-128_E-100.txt
|
| 43 |
+
β βββ 0utput_exp-mamba3_460h_W-512_D-30_S-16_B-64_E-50.txt
|
| 44 |
+
β βββ 0utput_exp-mamba3_enc_960h_W-320_D-48_S-16_b-64_E-100.txt
|
| 45 |
+
β βββ 0utput_exp-mamba_dt_bias_hier_460h_W-512_D-30_S-16.txt
|
| 46 |
+
β βββ 0utput_exp-v75_460h_hier_gating_256_42.txt
|
| 47 |
+
β βββ 0utput_exp-v77_960h_hier_gating_320_48.txt
|
| 48 |
+
β βββ mamba3_elm_training.log
|
| 49 |
+
β βββ mamba_elm_training.log
|
| 50 |
+
βββ .gitattributes
|
| 51 |
+
βββ README.md
|
| 52 |
+
```
|