Update README.md

README.md

@@ -1,3 +1,41 @@
----
-license: apache-2.0
----
+---
+license: apache-2.0
+base_model:
+- openai/whisper-base
+---
+
+# Whisper base
+
+Whisper base [model](https://huggingface.co/openai/whisper-base) converted to ONNX format for [onnx_asr](https://github.com/istupakov/onnx-asr).
+
+## Install onnx-asr
+```shell
+pip install onnx-asr[cpu,hub]
+```
+
+## Load whisper-base model and recognize wav file
+```py
+import onnx_asr
+model = onnx_asr.load_model("whisper-base")
+print(model.recognize("test.wav"))
+```
+
+## Code for models export
+
+Export Whisper to ONNX with `onnxruntime` ([whisper.convert_to_onnx](https://github.com/microsoft/onnxruntime/blob/main/onnxruntime/python/tools/transformers/models/whisper/README.md)).
+
+Download model and export with Beam Search and Forced Decoder Input Ids:
+```shell
+python3 -m onnxruntime.transformers.models.whisper.convert_to_onnx -m openai/whisper-base --output whisper-onnx --use_external_data_format --use_forced_decoder_ids --optimize_onnx --precision fp32
+```
+
+Save tokenizer vocabulary
+```py
+from transformers import WhisperTokenizer
+
+tokenizer = WhisperTokenizer.from_pretrained("openai/whisper-base")
+
+with open("whisper-onnx/vocab.txt", "w") as f:
+    for token, id in tokenizer.get_vocab().items():
+        f.write(f"{token} {id}\n")
+```