Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
Safetensors
Norwegian
whisper
whisper-event
norwegian
Eval Results (legacy)
Instructions to use NbAiLab/whisper-small-nob with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use NbAiLab/whisper-small-nob with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("automatic-speech-recognition", model="NbAiLab/whisper-small-nob")# Load model directly from transformers import AutoProcessor, AutoModelForSpeechSeq2Seq processor = AutoProcessor.from_pretrained("NbAiLab/whisper-small-nob") model = AutoModelForSpeechSeq2Seq.from_pretrained("NbAiLab/whisper-small-nob", device_map="auto") - Notebooks
- Google Colab
- Kaggle
first
Browse files
run.sh
CHANGED
|
@@ -8,7 +8,7 @@ deepspeed --master_port 29501 run_speech_recognition_seq2seq_streaming.py \
|
|
| 8 |
--model_index_name="Whisper Small Norwegian" \
|
| 9 |
--max_steps="50000" \
|
| 10 |
--output_dir="./" \
|
| 11 |
-
--per_device_train_batch_size="
|
| 12 |
--per_device_eval_batch_size="32" \
|
| 13 |
--gradient_accumulation_steps="1" \
|
| 14 |
--logging_steps="25" \
|
|
|
|
| 8 |
--model_index_name="Whisper Small Norwegian" \
|
| 9 |
--max_steps="50000" \
|
| 10 |
--output_dir="./" \
|
| 11 |
+
--per_device_train_batch_size="190" \
|
| 12 |
--per_device_eval_batch_size="32" \
|
| 13 |
--gradient_accumulation_steps="1" \
|
| 14 |
--logging_steps="25" \
|
runs/Dec15_13-01-19_dante/1671105709.2244825/events.out.tfevents.1671105709.dante.1804775.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:37b364a02fcf311f3dfb91e70c8fec3c510ffcec4b18fd6d230e19c735c20b88
|
| 3 |
+
size 5888
|
runs/Dec15_13-01-19_dante/events.out.tfevents.1671105709.dante.1804775.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:00eb2de2f8702cad22db3fae634f00a383fb2ec83afa84e5aa19eacfebaf4d64
|
| 3 |
+
size 4291
|