End of training

README.md

@@ -46,7 +46,7 @@ datasets:
     #   output:
 
 test_datasets:
-  - path: data/test.jsonl
+  - path: data/eval.jsonl
     ds_type: json
     # You need to specify a split. For "json" datasets the default split is called "train".
     split: train
@@ -116,7 +116,7 @@ xformers_attention: null
 
 This model is a fine-tuned version of [TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T](https://huggingface.co/TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T) on the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 3.3833
+- Loss: 1.5572
 
 ## Model description
 
@@ -148,10 +148,22 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 5.5347        | 1.0   | 1    | 3.4154          |
-| 6.0917        | 2.0   | 2    | 3.4136          |
-| 5.2963        | 2.0   | 3    | 3.4033          |
-| 7.2209        | 3.0   | 4    | 3.3833          |
+| 6.4934        | 0.25  | 1    | 2.0690          |
+| 2.5023        | 0.5   | 2    | 2.0673          |
+| 4.9022        | 0.75  | 3    | 2.0621          |
+| 5.6912        | 1.0   | 4    | 2.0491          |
+| 5.1317        | 1.25  | 5    | 2.0230          |
+| 5.5762        | 1.25  | 6    | 1.9738          |
+| 3.3504        | 1.5   | 7    | 1.9053          |
+| 5.1877        | 1.75  | 8    | 1.8346          |
+| 3.8815        | 2.0   | 9    | 1.7862          |
+| 3.5814        | 2.25  | 10   | 1.7475          |
+| 3.3579        | 2.25  | 11   | 1.6987          |
+| 3.5511        | 2.5   | 12   | 1.6555          |
+| 3.3339        | 2.75  | 13   | 1.6107          |
+| 2.8774        | 3.0   | 14   | 1.5778          |
+| 3.1427        | 3.25  | 15   | 1.5620          |
+| 3.3465        | 3.25  | 16   | 1.5572          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:046802b2fc28576bab178ad74beb832b2027a1a63fc31886d205489502db2442
+oid sha256:44772c4af3abc75d6063ca37102b982b62d41ac5fad308cde51f7d47e39986ef
 size 101036698