Upload WhisperForConditionalGeneration

config.json

@@ -0,0 +1,65 @@
+{
+  "_name_or_path": "./models/whisper-large-v3-no-numbers",
+  "activation_dropout": 0.0,
+  "activation_function": "gelu",
+  "apply_spec_augment": false,
+  "architectures": [
+    "WhisperForConditionalGeneration"
+  ],
+  "attention_dropout": 0.0,
+  "begin_suppress_tokens": [
+    210,
+    49839
+  ],
+  "bos_token_id": 49839,
+  "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,
+  "dropout": 0.0,
+  "encoder_attention_heads": 20,
+  "encoder_ffn_dim": 5120,
+  "encoder_layerdrop": 0.0,
+  "encoder_layers": 32,
+  "eos_token_id": 49839,
+  "forced_decoder_ids": [
+    [
+      1,
+      49841
+    ],
+    [
+      2,
+      49941
+    ],
+    [
+      3,
+      49945
+    ]
+  ],
+  "init_std": 0.02,
+  "is_encoder_decoder": true,
+  "mask_feature_length": 10,
+  "mask_feature_min_masks": 0,
+  "mask_feature_prob": 0.0,
+  "mask_time_length": 10,
+  "mask_time_min_masks": 2,
+  "mask_time_prob": 0.05,
+  "max_length": 448,
+  "max_source_positions": 1500,
+  "max_target_positions": 448,
+  "median_filter_width": 7,
+  "model_type": "whisper",
+  "num_hidden_layers": 32,
+  "num_mel_bins": 128,
+  "pad_token_id": 49838,
+  "scale_embedding": false,
+  "suppress_tokens": [],
+  "torch_dtype": "float32",
+  "transformers_version": "4.44.0",
+  "use_cache": true,
+  "use_weighted_layer_sum": false,
+  "vocab_size": 51447
+}

generation_config.json

@@ -0,0 +1,176 @@
+{
+  "alignment_heads": [
+    [
+      7,
+      0
+    ],
+    [
+      10,
+      17
+    ],
+    [
+      12,
+      18
+    ],
+    [
+      13,
+      12
+    ],
+    [
+      16,
+      1
+    ],
+    [
+      17,
+      14
+    ],
+    [
+      19,
+      11
+    ],
+    [
+      21,
+      4
+    ],
+    [
+      24,
+      1
+    ],
+    [
+      25,
+      6
+    ]
+  ],
+  "begin_suppress_tokens": [
+    210,
+    49839
+  ],
+  "bos_token_id": 49839,
+  "decoder_start_token_id": 49840,
+  "eos_token_id": 49839,
+  "forced_decoder_ids": [
+    [
+      1,
+      null
+    ],
+    [
+      2,
+      49941
+    ]
+  ],
+  "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
+  },
+  "max_initial_timestamp_index": 50,
+  "max_length": 448,
+  "no_timestamps_token_id": 49946,
+  "pad_token_id": 49839,
+  "prev_sot_token_id": 49944,
+  "return_timestamps": false,
+  "suppress_tokens": [],
+  "task_to_id": {
+    "transcribe": 49942,
+    "translate": 49941
+  },
+  "transformers_version": "4.44.0"
+}

model-00001-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6269c5b070d7b142a9ad236f3aa540d51e3513de94aa4ce2fb4cc3cabbb2e147
+size 4991303600

model-00002-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:630ca774672856d2e0e39a702e590f635a1cfc5726a64b6578ab46dd367369a9
+size 1180663192