lucienbaumgartner/mtg-spike-multilabel-distilbert

README.md

@@ -18,11 +18,11 @@ 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 the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.3550
-- F1 Micro: 0.3846
-- F1 Macro: 0.3832
-- F1 Weighted: 0.3836
-- Accuracy: 0.3008
+- Loss: 0.2704
+- F1 Micro: 0.6957
+- F1 Macro: 0.6938
+- F1 Weighted: 0.6942
+- Accuracy: 0.6288
 
 ## Model description
 
@@ -53,16 +53,16 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss | F1 Micro | F1 Macro | F1 Weighted | Accuracy |
 |:-------------:|:-----:|:----:|:---------------:|:--------:|:--------:|:-----------:|:--------:|
-| No log        | 1.0   | 288  | 0.4072          | 0.0078   | 0.0074   | 0.0077      | 0.0859   |
-| 0.4124        | 2.0   | 576  | 0.3695          | 0.2230   | 0.2164   | 0.2160      | 0.1914   |
-| 0.4124        | 3.0   | 864  | 0.3614          | 0.3195   | 0.3123   | 0.3118      | 0.2578   |
-| 0.3467        | 4.0   | 1152 | 0.3584          | 0.3580   | 0.3547   | 0.3539      | 0.2695   |
-| 0.3467        | 5.0   | 1440 | 0.3550          | 0.3846   | 0.3832   | 0.3836      | 0.3008   |
-| 0.3168        | 6.0   | 1728 | 0.3578          | 0.4127   | 0.4089   | 0.4091      | 0.3164   |
-| 0.291         | 7.0   | 2016 | 0.3618          | 0.4136   | 0.4096   | 0.4099      | 0.3164   |
-| 0.291         | 8.0   | 2304 | 0.3636          | 0.4175   | 0.4134   | 0.4137      | 0.3242   |
-| 0.281         | 9.0   | 2592 | 0.3640          | 0.4235   | 0.4197   | 0.4201      | 0.3320   |
-| 0.281         | 10.0  | 2880 | 0.3647          | 0.4125   | 0.4094   | 0.4096      | 0.3203   |
+| No log        | 1.0   | 406  | 0.3799          | 0.0741   | 0.0712   | 0.0715      | 0.1468   |
+| 0.4142        | 2.0   | 812  | 0.2767          | 0.5943   | 0.5891   | 0.5903      | 0.4986   |
+| 0.253         | 3.0   | 1218 | 0.2743          | 0.6603   | 0.6510   | 0.6523      | 0.5734   |
+| 0.2016        | 4.0   | 1624 | 0.2716          | 0.6656   | 0.6601   | 0.6611      | 0.5873   |
+| 0.1857        | 5.0   | 2030 | 0.2704          | 0.6957   | 0.6938   | 0.6942      | 0.6288   |
+| 0.1857        | 6.0   | 2436 | 0.2712          | 0.6862   | 0.6854   | 0.6858      | 0.6233   |
+| 0.1636        | 7.0   | 2842 | 0.2790          | 0.7011   | 0.6999   | 0.7004      | 0.6399   |
+| 0.1544        | 8.0   | 3248 | 0.2754          | 0.7068   | 0.7065   | 0.7068      | 0.6510   |
+| 0.1455        | 9.0   | 3654 | 0.2769          | 0.7087   | 0.7082   | 0.7085      | 0.6482   |
+| 0.1424        | 10.0  | 4060 | 0.2781          | 0.7096   | 0.7089   | 0.7092      | 0.6510   |
 
 
 ### Framework versions

adapter_config.json

@@ -24,8 +24,8 @@
   "revision": null,
   "target_modules": [
     "v_lin",
-    "k_lin",
-    "q_lin"
+    "q_lin",
+    "k_lin"
   ],
   "task_type": "SEQ_CLS",
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a6dae306d1f533dedc0eb27d9cd48b67d9e7eb8e488bc3bd35fde1e2be3f169c
+oid sha256:313120a73de8d836ed235ec930b762c6d4aee2fee14833b884a9a2956f9353d0
 size 3268052

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:eab9a001314ff0d32993e205c39ca97adb423ac57659d64ab8cd7eda9c0737fe
+oid sha256:f535e62a370b67527208c18429db1c62c22fbdd074794eca5d3033acfb068356
 size 4728