CuATR-distilbert-LoRA

README.md

@@ -18,9 +18,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.6934
-- Accuracy: 0.3913
-- F1: 0.3
+- Loss: 0.6944
+- Accuracy: 0.4348
+- F1: 0.0
 
 ## Model description
 
@@ -52,18 +52,18 @@ The following hyperparameters were used during training:
 
 ### Training results
 
-| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1     |
-|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|
-| 0.6945        | 0.67  | 1    | 0.6945          | 0.3478   | 0.2105 |
-| 0.6983        | 2.0   | 3    | 0.6942          | 0.3478   | 0.2105 |
-| 0.689         | 2.67  | 4    | 0.6941          | 0.3478   | 0.2105 |
-| 0.6901        | 4.0   | 6    | 0.6938          | 0.3478   | 0.2105 |
-| 0.6986        | 4.67  | 7    | 0.6938          | 0.3478   | 0.2105 |
-| 0.6937        | 6.0   | 9    | 0.6936          | 0.3478   | 0.2105 |
-| 0.6965        | 6.67  | 10   | 0.6936          | 0.3478   | 0.2105 |
-| 0.7016        | 8.0   | 12   | 0.6935          | 0.3913   | 0.3    |
-| 0.6946        | 8.67  | 13   | 0.6935          | 0.3913   | 0.3    |
-| 0.6871        | 9.33  | 14   | 0.6934          | 0.3913   | 0.3    |
+| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1  |
+|:-------------:|:-----:|:----:|:---------------:|:--------:|:---:|
+| 0.688         | 0.67  | 1    | 0.6963          | 0.4348   | 0.0 |
+| 0.6869        | 2.0   | 3    | 0.6957          | 0.4348   | 0.0 |
+| 0.699         | 2.67  | 4    | 0.6954          | 0.4348   | 0.0 |
+| 0.7055        | 4.0   | 6    | 0.6952          | 0.4348   | 0.0 |
+| 0.6813        | 4.67  | 7    | 0.6950          | 0.4348   | 0.0 |
+| 0.6985        | 6.0   | 9    | 0.6947          | 0.4348   | 0.0 |
+| 0.685         | 6.67  | 10   | 0.6946          | 0.4348   | 0.0 |
+| 0.6865        | 8.0   | 12   | 0.6945          | 0.4348   | 0.0 |
+| 0.6874        | 8.67  | 13   | 0.6944          | 0.4348   | 0.0 |
+| 0.6984        | 9.33  | 14   | 0.6944          | 0.4348   | 0.0 |
 
 
 ### Framework versions

adapter_config.json

@@ -9,7 +9,7 @@
   "layers_pattern": null,
   "layers_to_transform": null,
   "lora_alpha": 32,
-  "lora_dropout": 0.1,
+  "lora_dropout": 0.01,
   "modules_to_save": null,
   "peft_type": "LORA",
   "r": 16,
@@ -17,8 +17,8 @@
   "revision": null,
   "target_modules": [
     "k_lin",
-    "v_lin",
-    "q_lin"
+    "q_lin",
+    "v_lin"
   ],
   "task_type": "TOKEN_CLS"
 }

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b766422ece92960825408fb8493b46a0251e5519efd4857216af48c0630a1dc1
+oid sha256:225f1f3a331bb870164a5e01f9fc9dca50d9a5ff690342719ceb60bda49ff8e4
 size 1774720

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e7e165aba087bc2c3e63cae4b447f649f74f6b156aca546e343aa841a3938194
+oid sha256:1645f3a911c924d9a567624957f3db17e0efc0fcaeb4703c384105c6cf977b8b
 size 4600