End of training

README.md

@@ -30,8 +30,7 @@ load_in_4bit: true
 strict: false
 
 datasets:
-  - path: /tmp/fizzbuzz-ft/datasets
-    data_files: /tmp/fizzbuzz-ft/datasets/training-set-alpaca.json
+  - path: noeloco/fizzbuzz-sft
     type: alpaca
     ds_type: json
 
@@ -46,7 +45,7 @@ sequence_len: 2048
 sample_packing: false
 pad_to_sequence_len: true
 
-adapter: qlora
+adapter: lora
 lora_model_dir:
 lora_r: 16
 lora_alpha: 8
@@ -102,7 +101,7 @@ special_tokens:
 
 This model is a fine-tuned version of [codellama/CodeLlama-7b-hf](https://huggingface.co/codellama/CodeLlama-7b-hf) on the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.0295
+- Loss: 0.0210
 
 ## Model description
 
@@ -134,18 +133,18 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 2.0177        | 0.01  | 1    | 2.5549          |
-| 0.603         | 0.26  | 18   | 0.8667          |
-| 0.3026        | 0.51  | 36   | 0.2340          |
-| 0.0977        | 0.77  | 54   | 0.1274          |
-| 0.1101        | 1.03  | 72   | 0.1098          |
-| 0.0503        | 1.29  | 90   | 0.0469          |
-| 0.0753        | 1.54  | 108  | 0.0516          |
-| 0.2285        | 1.8   | 126  | 0.0192          |
-| 0.0647        | 2.06  | 144  | 0.0386          |
-| 0.0494        | 2.31  | 162  | 0.0334          |
-| 0.0552        | 2.57  | 180  | 0.0293          |
-| 0.0888        | 2.83  | 198  | 0.0295          |
+| 2.0829        | 0.01  | 1    | 2.5224          |
+| 0.6045        | 0.26  | 18   | 0.8178          |
+| 0.3357        | 0.51  | 36   | 0.2672          |
+| 0.1057        | 0.77  | 54   | 0.1210          |
+| 0.1046        | 1.03  | 72   | 0.0818          |
+| 0.052         | 1.29  | 90   | 0.0458          |
+| 0.0641        | 1.54  | 108  | 0.0363          |
+| 0.1952        | 1.8   | 126  | 0.0213          |
+| 0.0573        | 2.06  | 144  | 0.0362          |
+| 0.0346        | 2.31  | 162  | 0.0284          |
+| 0.0513        | 2.57  | 180  | 0.0221          |
+| 0.0865        | 2.83  | 198  | 0.0210          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4ca1b3f41f48bd83f6939570330d3b5133250530794bf42ad3cc23a91023705b
-size 80115914
+oid sha256:85bb4b3e876da33f2bbc71ffbf9c5a7955242c2d50961739c57a68990b807677
+size 160069834