End of training

README.md

@@ -18,10 +18,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [xlm-roberta-large](https://huggingface.co/xlm-roberta-large) on the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.6560
-- F1 Macro: 0.8840
-- F1 Weighted: 0.8867
-- Accuracy: 0.8867
+- Loss: 0.8214
+- Accuracy: 0.6439
+- F1 Macro: 0.6444
+- F1 Weighted: 0.6438
 
 ## Model description
 
@@ -40,10 +40,12 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 1e-05
+- learning_rate: 2e-05
 - train_batch_size: 16
 - eval_batch_size: 16
 - seed: 42
+- gradient_accumulation_steps: 2
+- total_train_batch_size: 32
 - optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
 - lr_scheduler_type: linear
 - lr_scheduler_warmup_ratio: 0.1
@@ -52,23 +54,20 @@ The following hyperparameters were used during training:
 
 ### Training results
 
-| Training Loss | Epoch | Step | Validation Loss | F1 Macro | F1 Weighted | Accuracy |
-|:-------------:|:-----:|:----:|:---------------:|:--------:|:-----------:|:--------:|
-| 0.6484        | 1.0   | 164  | 0.3818          | 0.8347   | 0.8402      | 0.8430   |
-| 0.3531        | 2.0   | 328  | 0.5118          | 0.8365   | 0.8377      | 0.8369   |
-| 0.2805        | 3.0   | 492  | 0.5171          | 0.8408   | 0.8423      | 0.8415   |
-| 0.1972        | 4.0   | 656  | 0.6697          | 0.8630   | 0.8650      | 0.8643   |
-| 0.1422        | 5.0   | 820  | 0.5505          | 0.8781   | 0.8816      | 0.8826   |
-| 0.0984        | 6.0   | 984  | 0.6658          | 0.8853   | 0.8875      | 0.8872   |
-| 0.0607        | 7.0   | 1148 | 0.8728          | 0.8834   | 0.8858      | 0.8857   |
-| 0.0483        | 8.0   | 1312 | 0.9191          | 0.8704   | 0.8725      | 0.8720   |
-| 0.0266        | 9.0   | 1476 | 0.9519          | 0.8670   | 0.8694      | 0.8689   |
-| 0.0298        | 10.0  | 1640 | 0.9161          | 0.8694   | 0.8721      | 0.8720   |
+| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 Macro | F1 Weighted |
+|:-------------:|:-----:|:----:|:---------------:|:--------:|:--------:|:-----------:|
+| 1.1168        | 1.0   | 121  | 0.9948          | 0.5421   | 0.4476   | 0.4564      |
+| 0.9225        | 2.0   | 242  | 0.8052          | 0.6410   | 0.6464   | 0.6474      |
+| 0.8084        | 3.0   | 363  | 0.8085          | 0.6961   | 0.6760   | 0.6807      |
+| 0.6813        | 4.0   | 484  | 0.7785          | 0.6670   | 0.6698   | 0.6676      |
+| 0.5894        | 5.0   | 605  | 0.8427          | 0.6681   | 0.6723   | 0.6709      |
+| 0.4791        | 6.0   | 726  | 0.9016          | 0.6722   | 0.6643   | 0.6663      |
+| 0.4063        | 7.0   | 847  | 0.9857          | 0.6712   | 0.6715   | 0.6732      |
 
 
 ### Framework versions
 
-- Transformers 4.57.1
-- Pytorch 2.8.0+cu126
-- Datasets 4.0.0
+- Transformers 4.57.3
+- Pytorch 2.9.0+cu126
+- Datasets 4.4.1
 - Tokenizers 0.22.1

config.json

@@ -10,8 +10,18 @@
   "hidden_act": "gelu",
   "hidden_dropout_prob": 0.1,
   "hidden_size": 1024,
+  "id2label": {
+    "0": "N",
+    "1": "NEU",
+    "2": "P"
+  },
   "initializer_range": 0.02,
   "intermediate_size": 4096,
+  "label2id": {
+    "N": 0,
+    "NEU": 1,
+    "P": 2
+  },
   "layer_norm_eps": 1e-05,
   "max_position_embeddings": 514,
   "model_type": "xlm-roberta",
@@ -20,8 +30,7 @@
   "output_past": true,
   "pad_token_id": 1,
   "position_embedding_type": "absolute",
-  "problem_type": "single_label_classification",
-  "transformers_version": "4.57.1",
+  "transformers_version": "4.57.3",
   "type_vocab_size": 1,
   "use_cache": true,
   "vocab_size": 250002

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:88acffeb20c81cecae797ad39b36fbc852d802a510d02e3d088d3279d24105f8
-size 2239618672
+oid sha256:2f8b249837d9e9353ee4e276bd1299c53b528da37769f2e0c48f5b644d7f33b8
+size 2239622772

tokenizer.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:93189c5d9a15db043017cfd920e00cf72fe9a4220bd74b460b635f6aa85a61a2
+oid sha256:3c088c06cf975b7097e469bd69630cdb0d675c6db1ce3af1042b6e19c6d01f22
 size 17082999

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:69c36a33a2bd1d4531d988e4fd71d43a45bc7ff721dfad4a5f2c3e34b3280a5e
+oid sha256:407e0060e33a5fde8c99602a34a54143913cdf1a562e8f6b2dcbbb107fb536c1
 size 5841