Remove additional tokens

README.md

@@ -1,13 +1,6 @@
 ---
-license: apache-2.0
-library_name: transformers
-pipeline_tag: automatic-speech-recognition
-tags:
-- asr
-- Pytorch
-- pruned
-- audio
-- automatic-speech-recognition
+base_model:
+- openai/whisper-large-v3
 language:
 - en
 - zh
@@ -108,8 +101,15 @@ language:
 - ba
 - jw
 - su
-base_model:
-- openai/whisper-large-v3
+library_name: transformers
+license: apache-2.0
+pipeline_tag: automatic-speech-recognition
+tags:
+- asr
+- Pytorch
+- pruned
+- audio
+- automatic-speech-recognition
 ---
 
 # Whisper-large-v3-no-numbers

config.json

@@ -9,34 +9,34 @@
   "attention_dropout": 0.0,
   "begin_suppress_tokens": [
     210,
-    49839
+    49826
   ],
-  "bos_token_id": 49839,
+  "bos_token_id": 49826,
   "classifier_proj_size": 256,
   "d_model": 1280,
   "decoder_attention_heads": 20,
   "decoder_ffn_dim": 5120,
   "decoder_layerdrop": 0.0,
   "decoder_layers": 32,
-  "decoder_start_token_id": 49840,
+  "decoder_start_token_id": 49827,
   "dropout": 0.0,
   "encoder_attention_heads": 20,
   "encoder_ffn_dim": 5120,
   "encoder_layerdrop": 0.0,
   "encoder_layers": 32,
-  "eos_token_id": 49839,
+  "eos_token_id": 49826,
   "forced_decoder_ids": [
     [
       1,
-      49841
+      49828
     ],
     [
       2,
-      49941
+      49928
     ],
     [
       3,
-      49945
+      49932
     ]
   ],
   "init_std": 0.02,
@@ -54,12 +54,12 @@
   "model_type": "whisper",
   "num_hidden_layers": 32,
   "num_mel_bins": 128,
-  "pad_token_id": 49838,
+  "pad_token_id": 49825,
   "scale_embedding": false,
   "suppress_tokens": [],
   "torch_dtype": "float32",
   "transformers_version": "4.44.0",
   "use_cache": true,
   "use_weighted_layer_sum": false,
-  "vocab_size": 51447
+  "vocab_size": 51434
 }

generation_config.json

@@ -43,11 +43,11 @@
   ],
   "begin_suppress_tokens": [
     210,
-    49839
+    49826
   ],
-  "bos_token_id": 49839,
-  "decoder_start_token_id": 49840,
-  "eos_token_id": 49839,
+  "bos_token_id": 49826,
+  "decoder_start_token_id": 49827,
+  "eos_token_id": 49826,
   "forced_decoder_ids": [
     [
       1,
@@ -55,122 +55,122 @@
     ],
     [
       2,
-      49941
+      49928
     ]
   ],
   "is_multilingual": true,
   "lang_to_id": {
-    "<|af|>": 49909,
-    "<|am|>": 49916,
-    "<|ar|>": 49854,
-    "<|as|>": 49932,
-    "<|az|>": 49886,
-    "<|ba|>": 49937,
-    "<|be|>": 49912,
-    "<|bg|>": 49874,
-    "<|bn|>": 49884,
-    "<|bo|>": 49929,
-    "<|br|>": 49891,
-    "<|bs|>": 49897,
-    "<|ca|>": 49852,
-    "<|cs|>": 49865,
-    "<|cy|>": 49879,
-    "<|da|>": 49867,
-    "<|de|>": 49843,
-    "<|el|>": 49863,
-    "<|en|>": 49841,
-    "<|es|>": 49844,
-    "<|et|>": 49889,
-    "<|eu|>": 49892,
-    "<|fa|>": 49882,
-    "<|fi|>": 49859,
-    "<|fo|>": 49920,
-    "<|fr|>": 49847,
-    "<|gl|>": 49901,
-    "<|gu|>": 49915,
-    "<|haw|>": 49934,
-    "<|ha|>": 49936,
-    "<|he|>": 49861,
-    "<|hi|>": 49858,
-    "<|hr|>": 49873,
-    "<|ht|>": 49921,
-    "<|hu|>": 49868,
-    "<|hy|>": 49894,
-    "<|id|>": 49857,
-    "<|is|>": 49893,
-    "<|it|>": 49856,
-    "<|ja|>": 49848,
-    "<|jw|>": 49938,
-    "<|ka|>": 49911,
-    "<|kk|>": 49898,
-    "<|km|>": 49905,
-    "<|kn|>": 49888,
-    "<|ko|>": 49846,
-    "<|la|>": 49876,
-    "<|lb|>": 49927,
-    "<|ln|>": 49935,
-    "<|lo|>": 49918,
-    "<|lt|>": 49875,
-    "<|lv|>": 49883,
-    "<|mg|>": 49931,
-    "<|mi|>": 49877,
-    "<|mk|>": 49890,
-    "<|ml|>": 49878,
-    "<|mn|>": 49896,
-    "<|mr|>": 49902,
-    "<|ms|>": 49864,
-    "<|mt|>": 49925,
-    "<|my|>": 49928,
-    "<|ne|>": 49895,
-    "<|nl|>": 49853,
-    "<|nn|>": 49924,
-    "<|no|>": 49870,
-    "<|oc|>": 49910,
-    "<|pa|>": 49903,
-    "<|pl|>": 49851,
-    "<|ps|>": 49922,
-    "<|pt|>": 49849,
-    "<|ro|>": 49866,
-    "<|ru|>": 49845,
-    "<|sa|>": 49926,
-    "<|sd|>": 49914,
-    "<|si|>": 49904,
-    "<|sk|>": 49880,
-    "<|sl|>": 49887,
-    "<|sn|>": 49906,
-    "<|so|>": 49908,
-    "<|sq|>": 49899,
-    "<|sr|>": 49885,
-    "<|su|>": 49939,
-    "<|sv|>": 49855,
-    "<|sw|>": 49900,
-    "<|ta|>": 49869,
-    "<|te|>": 49881,
-    "<|tg|>": 49913,
-    "<|th|>": 49871,
-    "<|tk|>": 49923,
-    "<|tl|>": 49930,
-    "<|tr|>": 49850,
-    "<|tt|>": 49933,
-    "<|uk|>": 49862,
-    "<|ur|>": 49872,
-    "<|uz|>": 49919,
-    "<|vi|>": 49860,
-    "<|yi|>": 49917,
-    "<|yo|>": 49907,
-    "<|yue|>": 49940,
-    "<|zh|>": 49842
+    "<|af|>": 49896,
+    "<|am|>": 49903,
+    "<|ar|>": 49841,
+    "<|as|>": 49919,
+    "<|az|>": 49873,
+    "<|ba|>": 49924,
+    "<|be|>": 49899,
+    "<|bg|>": 49861,
+    "<|bn|>": 49871,
+    "<|bo|>": 49916,
+    "<|br|>": 49878,
+    "<|bs|>": 49884,
+    "<|ca|>": 49839,
+    "<|cs|>": 49852,
+    "<|cy|>": 49866,
+    "<|da|>": 49854,
+    "<|de|>": 49830,
+    "<|el|>": 49850,
+    "<|en|>": 49828,
+    "<|es|>": 49831,
+    "<|et|>": 49876,
+    "<|eu|>": 49879,
+    "<|fa|>": 49869,
+    "<|fi|>": 49846,
+    "<|fo|>": 49907,
+    "<|fr|>": 49834,
+    "<|gl|>": 49888,
+    "<|gu|>": 49902,
+    "<|haw|>": 49921,
+    "<|ha|>": 49923,
+    "<|he|>": 49848,
+    "<|hi|>": 49845,
+    "<|hr|>": 49860,
+    "<|ht|>": 49908,
+    "<|hu|>": 49855,
+    "<|hy|>": 49881,
+    "<|id|>": 49844,
+    "<|is|>": 49880,
+    "<|it|>": 49843,
+    "<|ja|>": 49835,
+    "<|jw|>": 49925,
+    "<|ka|>": 49898,
+    "<|kk|>": 49885,
+    "<|km|>": 49892,
+    "<|kn|>": 49875,
+    "<|ko|>": 49833,
+    "<|la|>": 49863,
+    "<|lb|>": 49914,
+    "<|ln|>": 49922,
+    "<|lo|>": 49905,
+    "<|lt|>": 49862,
+    "<|lv|>": 49870,
+    "<|mg|>": 49918,
+    "<|mi|>": 49864,
+    "<|mk|>": 49877,
+    "<|ml|>": 49865,
+    "<|mn|>": 49883,
+    "<|mr|>": 49889,
+    "<|ms|>": 49851,
+    "<|mt|>": 49912,
+    "<|my|>": 49915,
+    "<|ne|>": 49882,
+    "<|nl|>": 49840,
+    "<|nn|>": 49911,
+    "<|no|>": 49857,
+    "<|oc|>": 49897,
+    "<|pa|>": 49890,
+    "<|pl|>": 49838,
+    "<|ps|>": 49909,
+    "<|pt|>": 49836,
+    "<|ro|>": 49853,
+    "<|ru|>": 49832,
+    "<|sa|>": 49913,
+    "<|sd|>": 49901,
+    "<|si|>": 49891,
+    "<|sk|>": 49867,
+    "<|sl|>": 49874,
+    "<|sn|>": 49893,
+    "<|so|>": 49895,
+    "<|sq|>": 49886,
+    "<|sr|>": 49872,
+    "<|su|>": 49926,
+    "<|sv|>": 49842,
+    "<|sw|>": 49887,
+    "<|ta|>": 49856,
+    "<|te|>": 49868,
+    "<|tg|>": 49900,
+    "<|th|>": 49858,
+    "<|tk|>": 49910,
+    "<|tl|>": 49917,
+    "<|tr|>": 49837,
+    "<|tt|>": 49920,
+    "<|uk|>": 49849,
+    "<|ur|>": 49859,
+    "<|uz|>": 49906,
+    "<|vi|>": 49847,
+    "<|yi|>": 49904,
+    "<|yo|>": 49894,
+    "<|yue|>": 49927,
+    "<|zh|>": 49829
   },
   "max_initial_timestamp_index": 50,
   "max_length": 448,
-  "no_timestamps_token_id": 49946,
-  "pad_token_id": 49839,
-  "prev_sot_token_id": 49944,
+  "no_timestamps_token_id": 49933,
+  "pad_token_id": 49826,
+  "prev_sot_token_id": 49931,
   "return_timestamps": false,
   "suppress_tokens": [],
   "task_to_id": {
-    "transcribe": 49942,
-    "translate": 49941
+    "transcribe": 49929,
+    "translate": 49928
   },
   "transformers_version": "4.44.0"
 }

model-00001-of-00002.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6269c5b070d7b142a9ad236f3aa540d51e3513de94aa4ce2fb4cc3cabbb2e147
-size 4991303600
+oid sha256:39e3f1a1f4283821eb92f0185c06d78ec703a342f350a2095731dbb7c8b2d215
+size 4991237040

model.safetensors.index.json

@@ -1,6 +1,6 @@
 {
   "metadata": {
-    "total_size": 6171816960
+    "total_size": 6171750400
   },
   "weight_map": {
     "model.decoder.embed_positions.weight": "model-00001-of-00002.safetensors",