Improve model accuracy

README.md

@@ -9,7 +9,7 @@ datasets:
 metrics:
 - accuracy
 model-index:
-- name: vit-base-patch16-224-in21k
+- name: vit-base-patch16-224-in21k-finetuned-cifar10
   results:
   - task:
       name: Image Classification
@@ -20,18 +20,18 @@ model-index:
     metrics:
     - name: Accuracy
       type: accuracy
-      value: 0.9793
+      value: 0.9877
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 should probably proofread and complete it, then remove this comment. -->
 
-# vit-base-patch16-224-in21k
+# vit-base-patch16-224-in21k-finetuned-cifar10
 
 This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the cifar-10 dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.3125
-- Accuracy: 0.9793
+- Loss: 0.1126
+- Accuracy: 0.9877
 
 ## Model description
 
@@ -41,15 +41,6 @@ More information needed
 
 More information needed
 
-## How to Use
-
-```Python
-from transformers import pipeline
-
-pipe = pipeline("image-classification", "avanishd/vit-base-patch16-224-in21k-finetuned-cifar10")
-pipe(image)
-```
-
 ## Training and evaluation data
 
 More information needed
@@ -68,18 +59,21 @@ The following hyperparameters were used during training:
 - optimizer: Use OptimizerNames.ADAMW_TORCH 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
-- num_epochs: 1
+- num_epochs: 3
+- mixed_precision_training: Native AMP
 
 ### Training results
 
-| Training Loss | Epoch  | Step | Validation Loss | Accuracy |
-|:-------------:|:------:|:----:|:---------------:|:--------:|
-| 0.5245        | 0.9984 | 312  | 0.3125          | 0.9793   |
+| Training Loss | Epoch | Step | Validation Loss | Accuracy |
+|:-------------:|:-----:|:----:|:---------------:|:--------:|
+| 0.4166        | 1.0   | 313  | 0.2324          | 0.9791   |
+| 0.3247        | 2.0   | 626  | 0.1320          | 0.9875   |
+| 0.2661        | 2.992 | 936  | 0.1126          | 0.9877   |
 
 
 ### Framework versions
 
-- Transformers 4.51.1
+- Transformers 4.51.3
 - Pytorch 2.6.0+cu124
 - Datasets 3.5.0
 - Tokenizers 0.21.1

config.json

@@ -45,5 +45,5 @@
   "problem_type": "single_label_classification",
   "qkv_bias": true,
   "torch_dtype": "float32",
-  "transformers_version": "4.51.1"
+  "transformers_version": "4.51.3"
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:910d503e2489adb08c94e2bd27cc8747fba2c45745ee0272ed9e3f51e99a50cb
+oid sha256:dbada918e6ecb8d9f2b92228584a9f1e7ba2f43d60cbac209f31943c1aaf0f2e
 size 343248584

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5d5a280e4071e79091e882cdc418ec344682d6a6a830c4f07e5104b7ac186ddb
-size 5304
+oid sha256:bc783bca3f8a6d71a5c78b14f8ecf55a4e76f3b44ebd32bf193d31c994715ba3
+size 5368