TheBloke/CodeLlama-7B-Instruct-AWQ-FaVe-rank32-2epochs-v2

README.md

@@ -18,7 +18,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [TheBloke/CodeLlama-7B-Instruct-AWQ](https://huggingface.co/TheBloke/CodeLlama-7B-Instruct-AWQ) on the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.4323
+- Loss: 0.4048
 
 ## Model description
 
@@ -52,13 +52,13 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch  | Step | Validation Loss |
 |:-------------:|:------:|:----:|:---------------:|
-| No log        | 0.2685 | 10   | 1.4509          |
-| 1.5667        | 0.5369 | 20   | 0.7012          |
-| 1.5667        | 0.8054 | 30   | 0.6017          |
-| 0.5581        | 1.0738 | 40   | 0.5068          |
-| 0.5581        | 1.3423 | 50   | 0.4670          |
-| 0.4108        | 1.6107 | 60   | 0.4462          |
-| 0.4108        | 1.8792 | 70   | 0.4323          |
+| No log        | 0.2685 | 10   | 1.3320          |
+| 1.5746        | 0.5369 | 20   | 0.6544          |
+| 1.5746        | 0.8054 | 30   | 0.5413          |
+| 0.5669        | 1.0738 | 40   | 0.4830          |
+| 0.5669        | 1.3423 | 50   | 0.4500          |
+| 0.44          | 1.6107 | 60   | 0.4254          |
+| 0.44          | 1.8792 | 70   | 0.4048          |
 
 
 ### Framework versions

adapter_config.json

@@ -20,8 +20,8 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "q_proj",
-    "v_proj"
+    "v_proj",
+    "q_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5cfd5d2c9119c85b745ea155ee95083135ab5de1a0ee7b04c9e79b1019babd69
+oid sha256:a16e3973574714d08902b5160b6b17d144e5435f6fd060aadb07fb87d9a68042
 size 67126104

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3440519dfedb7d40d27c25c94d82b1aa7dbc10caa7e22e77de4f1265c709f3d3
+oid sha256:3eedc157000916e57767649fca3afff535785ecafa05af0227a3a27dcf3a3c4a
 size 5112