irenewds/shawgpt-ft

README.md

@@ -16,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [TheBloke/Mistral-7B-Instruct-v0.2-GPTQ](https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.2-GPTQ) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 2.1943
+- Loss: 1.4926
 
 ## Model description
 
@@ -35,15 +35,15 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 0.0005
+- learning_rate: 0.0001
 - train_batch_size: 4
 - eval_batch_size: 4
 - seed: 42
-- gradient_accumulation_steps: 2
-- total_train_batch_size: 8
+- gradient_accumulation_steps: 4
+- total_train_batch_size: 16
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
-- lr_scheduler_warmup_steps: 100
+- lr_scheduler_warmup_steps: 2
 - num_epochs: 20
 - mixed_precision_training: Native AMP
 
@@ -51,31 +51,31 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch   | Step | Validation Loss |
 |:-------------:|:-------:|:----:|:---------------:|
-| 4.6274        | 0.9231  | 6    | 4.1674          |
-| 3.7807        | 2.0     | 13   | 3.8359          |
-| 3.9369        | 2.9231  | 19   | 3.3280          |
-| 2.7875        | 4.0     | 26   | 2.5993          |
-| 2.5085        | 4.9231  | 32   | 1.9752          |
-| 1.4919        | 6.0     | 39   | 1.4654          |
-| 1.3448        | 6.9231  | 45   | 1.3415          |
-| 1.0855        | 8.0     | 52   | 1.3057          |
-| 1.1615        | 8.9231  | 58   | 1.2917          |
-| 0.8948        | 10.0    | 65   | 1.2995          |
-| 0.9723        | 10.9231 | 71   | 1.3292          |
-| 0.7141        | 12.0    | 78   | 1.3735          |
-| 0.733         | 12.9231 | 84   | 1.4165          |
-| 0.5298        | 14.0    | 91   | 1.5106          |
-| 0.529         | 14.9231 | 97   | 1.7135          |
-| 0.3745        | 16.0    | 104  | 1.8745          |
-| 0.3654        | 16.9231 | 110  | 1.9549          |
-| 0.2598        | 18.0    | 117  | 2.1665          |
-| 0.2422        | 18.4615 | 120  | 2.1943          |
+| 4.6196        | 0.9231  | 3    | 4.0996          |
+| 4.3186        | 1.8462  | 6    | 3.8038          |
+| 3.9621        | 2.7692  | 9    | 3.5173          |
+| 2.7151        | 4.0     | 13   | 3.1519          |
+| 3.3264        | 4.9231  | 16   | 2.9077          |
+| 3.0314        | 5.8462  | 19   | 2.6983          |
+| 2.7911        | 6.7692  | 22   | 2.5124          |
+| 1.9218        | 8.0     | 26   | 2.2823          |
+| 2.3602        | 8.9231  | 29   | 2.1229          |
+| 2.1297        | 9.8462  | 32   | 1.9649          |
+| 1.9877        | 10.7692 | 35   | 1.8543          |
+| 1.3751        | 12.0    | 39   | 1.7460          |
+| 1.7619        | 12.9231 | 42   | 1.6779          |
+| 1.6587        | 13.8462 | 45   | 1.6203          |
+| 1.6074        | 14.7692 | 48   | 1.5695          |
+| 1.1674        | 16.0    | 52   | 1.5241          |
+| 1.503         | 16.9231 | 55   | 1.5053          |
+| 1.4711        | 17.8462 | 58   | 1.4951          |
+| 1.0311        | 18.4615 | 60   | 1.4926          |
 
 
 ### Framework versions
 
 - PEFT 0.13.2
 - Transformers 4.44.2
-- Pytorch 2.4.1+cu121
+- Pytorch 2.5.0+cu124
 - Datasets 3.0.2
 - Tokenizers 0.19.1

runs/Oct24_01-19-49_89957c487371/events.out.tfevents.1729732794.89957c487371.1061.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e345e90f096b96c359e2d1c8bf667450c53b7de5c0af12e7921a89ee739712e1
+size 14917

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9ecb0b04ff97632084a4a209ce86b9013f9a3a08c4053dd3faf2291be9feaf03
+oid sha256:5d24572c9554940e1f92edd78e10243cb6873325ac8bd219df60a27e7683379a
 size 5176