C

config.json

@@ -1,5 +1,4 @@
 {
-  "_name_or_path": "marma/wav2vec2-large-swedish-bart-base",
   "architectures": [
     "SpeechEncoderDecoderModel"
   ],
@@ -88,22 +87,22 @@
   },
   "decoder_start_token_id": 0,
   "encoder": {
-    "_name_or_path": "KBLab/wav2vec2-large-voxrex-swedish",
-    "activation_dropout": 0.05,
+    "_name_or_path": "KBLab/wav2vec2-large-voxrex",
+    "activation_dropout": 0.0,
     "adapter_kernel_size": 3,
     "adapter_stride": 2,
-    "add_adapter": false,
+    "add_adapter": true,
     "add_cross_attention": false,
     "apply_spec_augment": true,
     "architectures": [
-      "Wav2Vec2ForCTC"
+      "Wav2Vec2ForPreTraining"
     ],
     "attention_dropout": 0.1,
     "bad_words_ids": null,
     "bos_token_id": 1,
     "chunk_size_feed_forward": 0,
     "classifier_proj_size": 256,
-    "codevector_dim": 256,
+    "codevector_dim": 768,
     "contrastive_logits_temperature": 0.1,
     "conv_bias": true,
     "conv_dim": [
@@ -134,8 +133,8 @@
       2
     ],
     "cross_attention_hidden_size": null,
-    "ctc_loss_reduction": "mean",
-    "ctc_zero_infinity": true,
+    "ctc_loss_reduction": "sum",
+    "ctc_zero_infinity": false,
     "decoder_start_token_id": null,
     "diversity_loss_weight": 0.1,
     "diversity_penalty": 0.0,
@@ -147,14 +146,15 @@
     "feat_extract_activation": "gelu",
     "feat_extract_dropout": 0.0,
     "feat_extract_norm": "layer",
-    "feat_proj_dropout": 0.05,
+    "feat_proj_dropout": 0.0,
     "feat_quantizer_dropout": 0.0,
     "final_dropout": 0.0,
     "finetuning_task": null,
     "forced_bos_token_id": null,
     "forced_eos_token_id": null,
+    "gradient_checkpointing": false,
     "hidden_act": "gelu",
-    "hidden_dropout": 0.05,
+    "hidden_dropout": 0.1,
     "hidden_size": 1024,
     "id2label": {
       "0": "LABEL_0",
@@ -169,7 +169,7 @@
       "LABEL_1": 1
     },
     "layer_norm_eps": 1e-05,
-    "layerdrop": 0.05,
+    "layerdrop": 0.0,
     "length_penalty": 1.0,
     "mask_channel_length": 10,
     "mask_channel_min_space": 1,
@@ -183,7 +183,7 @@
     "mask_time_min_masks": 2,
     "mask_time_min_space": 1,
     "mask_time_other": 0.0,
-    "mask_time_prob": 0.05,
+    "mask_time_prob": 0.0,
     "mask_time_selection": "static",
     "max_length": 20,
     "min_length": 0,
@@ -208,7 +208,7 @@
     "pad_token_id": 0,
     "prefix": null,
     "problem_type": null,
-    "proj_codevector_dim": 256,
+    "proj_codevector_dim": 768,
     "pruned_heads": {},
     "remove_invalid_values": false,
     "repetition_penalty": 1.0,
@@ -248,19 +248,17 @@
     "transformers_version": "4.16.0.dev0",
     "use_bfloat16": false,
     "use_weighted_layer_sum": false,
-    "vocab_size": 46,
+    "vocab_size": 32,
     "xvector_output_dim": 512
   },
-  "encoder_add_adapter": true,
-  "encoder_feat_proj_dropout": 0.0,
-  "encoder_layerdrop": 0.0,
   "eos_token_id": 2,
   "is_encoder_decoder": true,
-  "max_length": 200,
+  "max_length": 40,
   "model_type": "speech-encoder-decoder",
-  "num_beams": 5,
   "pad_token_id": 1,
+  "processor_class": "Wav2Vec2Processor",
   "tie_word_embeddings": false,
   "torch_dtype": "float32",
-  "transformers_version": null
+  "transformers_version": null,
+  "use_cache": false
 }

preprocessor_config.json

@@ -1,4 +1,5 @@
 {
+  "_processor_class": null,
   "do_normalize": true,
   "feature_extractor_type": "Wav2Vec2FeatureExtractor",
   "feature_size": 1,

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f1591efa33cd05727ab96974eb5b4c1718ef40b94f0e0f8455fa51d82294582e
-size 1649302897
+oid sha256:4a6fb5a66ede6750cf6e94c6360fc4840d4363847a89a69a7a84369c6a549e2b
+size 1649294705

tokenizer_config.json

@@ -1 +1 @@
-{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "pad_token": "<pad>", "mask_token": "<mask>", "cls_token": "</s>", "sep_token": "</s>", "special_tokens_map_file": "/home/marmal/.cache/huggingface/transformers/eeba8b7f723def8f95951bb301a16f684eb8a2b6bfbc13e715d200a2df8a7e61.4405f81af4d0c9b936ce3368cc07277d6d61ec031eac1137ba8e4d9acedd61ee", "name_or_path": "KBLab/bart-base-swedish-cased", "tokenizer_class": "PreTrainedTokenizerFast"}
+{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "pad_token": "<pad>", "mask_token": "<mask>", "cls_token": "</s>", "sep_token": "</s>", "special_tokens_map_file": null, "name_or_path": "KBLab/bart-base-swedish-cased", "tokenizer_class": "PreTrainedTokenizerFast"}