up

README.md

@@ -0,0 +1,11 @@
+This is an example of how a kenLM model can be downloaded with [PyCTCDecode](https://github.com/kensho-technologies/pyctcdecode) .
+
+Simply run the following code:
+
+```python
+from pyctcdecode import BeamSearchDecoderCTC
+
+decoder = BeamSearchDecoderCTC.load_from_hf_hub("kensho/beamsearch_decoder_dummy")
+```
+
+The model was created by [Patrick von Platen](https://huggingface.co/patrickvonplaten) for demonstration purposes.

alphabet.json

@@ -0,0 +1 @@
+{"is_bpe": false, "labels": ["<unk>", "<s>", "</s>", "bugs", "bunny"]}

language_model/attrs.json

@@ -0,0 +1 @@
+{"alpha": 0.5, "beta": 1.5, "unk_score_offset": -10.0, "score_boundary": true}

language_model/bugs_bunny_kenlm.arpa

@@ -0,0 +1,19 @@
+\data\
+ngram 1=5
+ngram 2=5
+
+\1-grams:
+-10	<unk>	0
+0	<s>	0
+0	</s>	0
+0	bugs	0
+0	bunny	0
+
+\2-grams:
+0	bunny </s>
+-10	bugs </s>
+0	<s> bugs
+-10	<s> bunny
+0	bugs bunny
+
+\end\

language_model/unigrams.txt

@@ -0,0 +1,5 @@
+<unk>
+<s>
+</s>
+bugs
+bunny