Ai-Marshal/12.5k_A100_5epoch_adapter

README.md

@@ -20,7 +20,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [mistralai/Mixtral-8x7B-Instruct-v0.1](https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1) on the generator dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.5052
+- Loss: 0.5049
 
 ## Model description
 
@@ -52,11 +52,11 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 0.4994        | 1.0   | 559  | 0.5309          |
-| 0.5643        | 2.0   | 1118 | 0.5145          |
-| 0.5366        | 3.0   | 1677 | 0.5087          |
-| 0.4885        | 4.0   | 2236 | 0.5059          |
-| 0.489         | 5.0   | 2795 | 0.5052          |
+| 0.5879        | 1.0   | 559  | 0.5304          |
+| 0.5272        | 2.0   | 1118 | 0.5142          |
+| 0.5564        | 3.0   | 1677 | 0.5083          |
+| 0.5207        | 4.0   | 2236 | 0.5057          |
+| 0.5123        | 5.0   | 2795 | 0.5049          |
 
 
 ### Framework versions

adapter_config.json

@@ -19,13 +19,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
+    "lm_head",
     "k_proj",
     "o_proj",
-    "q_proj",
     "up_proj",
-    "v_proj",
-    "lm_head",
     "down_proj",
+    "v_proj",
+    "q_proj",
     "gate_proj"
   ],
   "task_type": "CAUSAL_LM",

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6e23de478b88b30d97416b23ab7b3f1210371fd0a7ca3c63be82f82bb87d49d1
+oid sha256:685cb2b722bc04b4f8ddc062061a6e7f943745a8cc49381d7b521d8eab8a281d
 size 581159464

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0bfd2215035f93c2441bdf425a24608949c5a667395d9c8b1e60c608c591928d
+oid sha256:b8bb05001bdd74040f72341ff81896ea930af295a4dee9218828ce6e3d84218d
 size 4920