End of training

README.md

@@ -1,7 +1,7 @@
 ---
 library_name: transformers
 license: apache-2.0
-base_model: allenai/led-base-16384
+base_model: stefanbschneider/led-base-16384-lfqa-ans-len-512
 tags:
 - generated_from_trainer
 model-index:
@@ -14,10 +14,11 @@ should probably proofread and complete it, then remove this comment. -->
 
 # led-base-16384-lfqa
 
-This model is a fine-tuned version of [allenai/led-base-16384](https://huggingface.co/allenai/led-base-16384) on an unknown dataset.
+This model is a fine-tuned version of [stefanbschneider/led-base-16384-lfqa-ans-len-512](https://huggingface.co/stefanbschneider/led-base-16384-lfqa-ans-len-512) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 3.6340
-- Rouge2: 0.027
+- Loss: 3.2615
+- Rouge2: 0.0416
+- Task: {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'}
 
 ## Model description
 
@@ -40,21 +41,45 @@ The following hyperparameters were used during training:
 - train_batch_size: 2
 - eval_batch_size: 2
 - seed: 42
-- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
+- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
 - lr_scheduler_type: linear
 - num_epochs: 1
+- mixed_precision_training: Native AMP
 
 ### Training results
 
-| Training Loss | Epoch  | Step | Validation Loss | Rouge2 |
-|:-------------:|:------:|:----:|:---------------:|:------:|
-| 3.895         | 0.3906 | 50   | 3.7506          | 0.0266 |
-| 3.6803        | 0.7812 | 100  | 3.6340          | 0.027  |
+| Training Loss | Epoch  | Step  | Validation Loss | Rouge2 | Task                                                                               |
+|:-------------:|:------:|:-----:|:---------------:|:------:|:----------------------------------------------------------------------------------:|
+| 3.4849        | 0.0395 | 2000  | 3.4233          | 0.0387 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.4744        | 0.0789 | 4000  | 3.4411          | 0.0398 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.4919        | 0.1184 | 6000  | 3.4251          | 0.0378 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.487         | 0.1578 | 8000  | 3.4200          | 0.0397 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.4443        | 0.1973 | 10000 | 3.3870          | 0.0376 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.4597        | 0.2367 | 12000 | 3.3914          | 0.0405 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.4525        | 0.2762 | 14000 | 3.3845          | 0.0398 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.4618        | 0.3156 | 16000 | 3.3752          | 0.0424 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.4573        | 0.3551 | 18000 | 3.3693          | 0.0421 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.4164        | 0.3945 | 20000 | 3.3640          | 0.042  | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.4125        | 0.4340 | 22000 | 3.3544          | 0.0412 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.3828        | 0.4734 | 24000 | 3.3423          | 0.0409 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.3965        | 0.5129 | 26000 | 3.3436          | 0.0416 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.3993        | 0.5524 | 28000 | 3.3339          | 0.0384 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.3909        | 0.5918 | 30000 | 3.3122          | 0.0414 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.3745        | 0.6313 | 32000 | 3.3158          | 0.0416 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.3665        | 0.6707 | 34000 | 3.3038          | 0.0424 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.3351        | 0.7102 | 36000 | 3.2915          | 0.0435 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.3629        | 0.7496 | 38000 | 3.2955          | 0.0436 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.3465        | 0.7891 | 40000 | 3.2888          | 0.0395 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.3127        | 0.8285 | 42000 | 3.2800          | 0.0414 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.3385        | 0.8680 | 44000 | 3.2767          | 0.0413 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.2882        | 0.9074 | 46000 | 3.2685          | 0.0437 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.3162        | 0.9469 | 48000 | 3.2639          | 0.0412 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
+| 3.3072        | 0.9863 | 50000 | 3.2615          | 0.0416 | {'name': 'Sequence-to-sequence Language Modeling', 'type': 'text2text-generation'} |
 
 
 ### Framework versions
 
-- Transformers 4.48.1
-- Pytorch 2.5.1
+- Transformers 4.48.3
+- Pytorch 2.5.1+cu121
 - Datasets 3.2.0
 - Tokenizers 0.21.0

generation_config.json

@@ -3,9 +3,9 @@
   "decoder_start_token_id": 0,
   "early_stopping": true,
   "length_penalty": 2.0,
-  "max_length": 1024,
+  "max_length": 512,
   "min_length": 100,
   "no_repeat_ngram_size": 3,
   "num_beams": 4,
-  "transformers_version": "4.48.1"
+  "transformers_version": "4.48.3"
 }