Samuael/amBART_261

README.md

@@ -1,5 +1,5 @@
 ---
-base_model: Samuael/amBART
+base_model: Samuael/amBART_261
 tags:
 - generated_from_trainer
 model-index:
@@ -12,17 +12,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # amBART
 
-This model is a fine-tuned version of [Samuael/amBART](https://huggingface.co/Samuael/amBART) on an unknown dataset.
-It achieves the following results on the evaluation set:
-- eval_loss: 3.4340
-- eval_wer: 0.9966
-- eval_cer: 0.8055
-- eval_bleu: 0.1833
-- eval_runtime: 8.1414
-- eval_samples_per_second: 44.096
-- eval_steps_per_second: 0.246
-- epoch: 11.0
-- step: 3905
+This model is a fine-tuned version of [Samuael/amBART_261](https://huggingface.co/Samuael/amBART_261) on an unknown dataset.
 
 ## Model description
 
@@ -41,13 +31,14 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 0.0002
-- train_batch_size: 256
-- eval_batch_size: 256
+- learning_rate: 0.02
+- train_batch_size: 128
+- eval_batch_size: 128
 - seed: 42
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
 - num_epochs: 50
+- mixed_precision_training: Native AMP
 
 ### Framework versions
 

config.json

@@ -1,22 +1,22 @@
 {
-  "_name_or_path": "Samuael/amBART",
-  "activation_dropout": 0.0,
+  "_name_or_path": "Samuael/amBART_261",
+  "activation_dropout": 0.05,
   "activation_function": "gelu",
   "architectures": [
     "MBartForConditionalGeneration"
   ],
-  "attention_dropout": 0.0,
+  "attention_dropout": 0.05,
   "bos_token_id": 0,
-  "classifier_dropout": 0.0,
+  "classifier_dropout": 0.01,
   "d_model": 512,
-  "decoder_attention_heads": 16,
+  "decoder_attention_heads": 8,
   "decoder_ffn_dim": 2048,
   "decoder_layerdrop": 0.01,
   "decoder_layers": 6,
   "dropout": 0.1,
-  "encoder_attention_heads": 16,
+  "encoder_attention_heads": 8,
   "encoder_ffn_dim": 2048,
-  "encoder_layerdrop": 0.0,
+  "encoder_layerdrop": 0.1,
   "encoder_layers": 6,
   "eos_token_id": 2,
   "forced_eos_token_id": 2,
@@ -31,5 +31,5 @@
   "torch_dtype": "float32",
   "transformers_version": "4.38.2",
   "use_cache": true,
-  "vocab_size": 1027
+  "vocab_size": 261
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e48b1ce78fe21170b4789afb20904bfd8f3b41f68909381b4fb2c806011a1255
-size 180813204
+oid sha256:25f1ced3bfe39cb729972f0794233e0ad53a3de7cf589966d19e92867e4cf9fa
+size 179241372

sentencepiece.bpe.model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8b2c9b3910462e773a7ace377cb26a40053dcfe4054d8eb8ecc5460495a93d9b
-size 253575
+oid sha256:a801c63cf0822cc3a880177fd5895196337d7e3813edde88c428061c263354a4
+size 240461

special_tokens_map.json

@@ -25,8 +25,7 @@
     "tr_TR",
     "vi_VN",
     "zh_CN",
-    "▁በኋላ",
-    "ብር"
+    "ቊ"
   ],
   "bos_token": {
     "content": "<s>",

tokenizer_config.json

@@ -33,7 +33,7 @@
       "special": true
     },
     "234": {
-      "content": "▁በኋላ",
+      "content": "ቊ",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -41,14 +41,6 @@
       "special": true
     },
     "235": {
-      "content": "ብር",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "1001": {
       "content": "ar_AR",
       "lstrip": false,
       "normalized": false,
@@ -56,7 +48,7 @@
       "single_word": false,
       "special": true
     },
-    "1002": {
+    "236": {
       "content": "cs_CZ",
       "lstrip": false,
       "normalized": false,
@@ -64,7 +56,7 @@
       "single_word": false,
       "special": true
     },
-    "1003": {
+    "237": {
       "content": "de_DE",
       "lstrip": false,
       "normalized": false,
@@ -72,7 +64,7 @@
       "single_word": false,
       "special": true
     },
-    "1004": {
+    "238": {
       "content": "en_XX",
       "lstrip": false,
       "normalized": false,
@@ -80,7 +72,7 @@
       "single_word": false,
       "special": true
     },
-    "1005": {
+    "239": {
       "content": "es_XX",
       "lstrip": false,
       "normalized": false,
@@ -88,7 +80,7 @@
       "single_word": false,
       "special": true
     },
-    "1006": {
+    "240": {
       "content": "et_EE",
       "lstrip": false,
       "normalized": false,
@@ -96,7 +88,7 @@
       "single_word": false,
       "special": true
     },
-    "1007": {
+    "241": {
       "content": "fi_FI",
       "lstrip": false,
       "normalized": false,
@@ -104,7 +96,7 @@
       "single_word": false,
       "special": true
     },
-    "1008": {
+    "242": {
       "content": "fr_XX",
       "lstrip": false,
       "normalized": false,
@@ -112,7 +104,7 @@
       "single_word": false,
       "special": true
     },
-    "1009": {
+    "243": {
       "content": "gu_IN",
       "lstrip": false,
       "normalized": false,
@@ -120,7 +112,7 @@
       "single_word": false,
       "special": true
     },
-    "1010": {
+    "244": {
       "content": "hi_IN",
       "lstrip": false,
       "normalized": false,
@@ -128,7 +120,7 @@
       "single_word": false,
       "special": true
     },
-    "1011": {
+    "245": {
       "content": "it_IT",
       "lstrip": false,
       "normalized": false,
@@ -136,7 +128,7 @@
       "single_word": false,
       "special": true
     },
-    "1012": {
+    "246": {
       "content": "ja_XX",
       "lstrip": false,
       "normalized": false,
@@ -144,7 +136,7 @@
       "single_word": false,
       "special": true
     },
-    "1013": {
+    "247": {
       "content": "kk_KZ",
       "lstrip": false,
       "normalized": false,
@@ -152,7 +144,7 @@
       "single_word": false,
       "special": true
     },
-    "1014": {
+    "248": {
       "content": "ko_KR",
       "lstrip": false,
       "normalized": false,
@@ -160,7 +152,7 @@
       "single_word": false,
       "special": true
     },
-    "1015": {
+    "249": {
       "content": "lt_LT",
       "lstrip": false,
       "normalized": false,
@@ -168,7 +160,7 @@
       "single_word": false,
       "special": true
     },
-    "1016": {
+    "250": {
       "content": "lv_LV",
       "lstrip": false,
       "normalized": false,
@@ -176,7 +168,7 @@
       "single_word": false,
       "special": true
     },
-    "1017": {
+    "251": {
       "content": "my_MM",
       "lstrip": false,
       "normalized": false,
@@ -184,7 +176,7 @@
       "single_word": false,
       "special": true
     },
-    "1018": {
+    "252": {
       "content": "ne_NP",
       "lstrip": false,
       "normalized": false,
@@ -192,7 +184,7 @@
       "single_word": false,
       "special": true
     },
-    "1019": {
+    "253": {
       "content": "nl_XX",
       "lstrip": false,
       "normalized": false,
@@ -200,7 +192,7 @@
       "single_word": false,
       "special": true
     },
-    "1020": {
+    "254": {
       "content": "ro_RO",
       "lstrip": false,
       "normalized": false,
@@ -208,7 +200,7 @@
       "single_word": false,
       "special": true
     },
-    "1021": {
+    "255": {
       "content": "ru_RU",
       "lstrip": false,
       "normalized": false,
@@ -216,7 +208,7 @@
       "single_word": false,
       "special": true
     },
-    "1022": {
+    "256": {
       "content": "si_LK",
       "lstrip": false,
       "normalized": false,
@@ -224,7 +216,7 @@
       "single_word": false,
       "special": true
     },
-    "1023": {
+    "257": {
       "content": "tr_TR",
       "lstrip": false,
       "normalized": false,
@@ -232,7 +224,7 @@
       "single_word": false,
       "special": true
     },
-    "1024": {
+    "258": {
       "content": "vi_VN",
       "lstrip": false,
       "normalized": false,
@@ -240,7 +232,7 @@
       "single_word": false,
       "special": true
     },
-    "1025": {
+    "259": {
       "content": "zh_CN",
       "lstrip": false,
       "normalized": false,
@@ -275,8 +267,7 @@
     "tr_TR",
     "vi_VN",
     "zh_CN",
-    "▁በኋላ",
-    "ብር"
+    "ቊ"
   ],
   "bos_token": "<s>",
   "clean_up_tokenization_spaces": true,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:89e3c4eea5f7f0e97067c89e15fc1896956960907cf92ea9ab7b5f40a6a93afe
+oid sha256:92e512ac163487c52b68e44df6a966d14898c0b664954e51fcdbb52e6e0c5a78
 size 4984