Model save

README.md

@@ -15,8 +15,8 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [microsoft/trocr-large-printed](https://huggingface.co/microsoft/trocr-large-printed) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.5153
-- Cer: 0.0312
+- Loss: 0.2391
+- Cer: 0.0033
 
 ## Model description
 
@@ -47,13 +47,13 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss | Cer    |
 |:-------------:|:-----:|:----:|:---------------:|:------:|
-| 2.0108        | 1.0   | 61   | 1.1284          | 0.1530 |
-| 0.5665        | 2.0   | 122  | 0.5153          | 0.0312 |
+| 0.5455        | 1.0   | 841  | 0.4618          | 0.0201 |
+| 0.2068        | 2.0   | 1682 | 0.2391          | 0.0033 |
 
 
 ### Framework versions
 
 - Transformers 4.45.2
-- Pytorch 2.5.1+cu121
+- Pytorch 2.1.0+cu118
 - Datasets 3.2.0
 - Tokenizers 0.20.3

generation_config.json

@@ -4,7 +4,7 @@
   "early_stopping": true,
   "eos_token_id": 2,
   "length_penalty": 0,
-  "max_length": 39,
+  "max_length": 46,
   "no_repeat_ngram_size": 3,
   "num_beams": 8,
   "pad_token_id": 1,