Update README.md
Browse files
README.md
CHANGED
|
@@ -10,3 +10,15 @@ Fine-tuned version of [`inesc-id/WhisperLv3-FT`](https://huggingface.co/inesc-id
|
|
| 10 |
- **Compute type:** float16
|
| 11 |
|
| 12 |
> **Note:** A paper with full training methodology, evaluation results, and benchmarks is currently under preparation and will be published soon.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
- **Compute type:** float16
|
| 11 |
|
| 12 |
> **Note:** A paper with full training methodology, evaluation results, and benchmarks is currently under preparation and will be published soon.
|
| 13 |
+
|
| 14 |
+
**Current results:**
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
| # | Modelo | WER (%) | CER (%) | RTF | Tempo (s) | RAM | eurospeech_clean | falabracarense_clean | MLS |
|
| 18 |
+
|---| :--- | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: |
|
| 19 |
+
| 1 | Neo-X-Euro-Bracarense(transformers) | 13.67% | 10.16% | 0.000 | 1208.3s | 0MB | 28.1% | 7.1% | 5.8% |
|
| 20 |
+
| 2 | EP-X(Faster-Whisper) | 18.93% | 14.56% | 0.000 | 1259.1s | 0MB | 42.5% | 7.6% | 6.6% |
|
| 21 |
+
| 3 | whisper-large-v3 | 26.99% | 19.57% | 0.000 | 1202.9s | 0MB | 41.8% | 33.9% | 5.2% |
|
| 22 |
+
| 4 | Nvidia-Canary-1b-v2 | 32.07% | 22.06% | 0.000 | 2296.4s | 0MB | 45.6% | 43.5% | 7.1% |
|
| 23 |
+
| | **MÉDIA (4 modelos c/ 3 datasets)** | **22.91%** | **16.58%** | | | | | | |
|
| 24 |
+
|